Search asymmetry for expansion and contraction motion in pigeons(Summary of Awarded Presentation at the 30th Annual Meeting)

(1)

The Japanese Psychonomic Society

NII-Electronic Library Service

TheJapanesePsychonomic Society

TheJapaneseJournatofPsychonomicS[ien[e

2012,VeL31,Ne.],S7-S8

Summary

of

Awarded

Presentation2G06

Searchasymmetry

for

expansion

and

Noriyuki

NAKAMuRA*

and

CenterforFrontier

Science,

Chiba

Uhiversity*,

contraction

motion

in

pigeons

]

Masako

_JiTsuMoRi*2

Flaculty

ofLetters,

Chiba

Uhiversity*2

We

examined whether

_pigeons

demonstrated

a search asymmetry

between

expansion and contraction

targets,

using a rotating

logarithrnic

spiral

_pattern

that

induces

illusory

in

humans

without

changes

in

the

_pattern's

_physical

size.

Four

_pigeons

were arbitraril},

divided

into

two

_groups.

Two

pigeons

in

the

Thr-get-Expansion

Group

were

trained

to

sear[h

for

1

expanding spiral

{target)

among

5

contracting ones

(distractors)

in

displays

cornprising a

total

of

6

spirals

(display

size=6).

TLvo

pigeons

in

the

Tlarget-Contraction

Group

were

trained

to search

for

1

[ontracting spiral among

5

expanding ones.

Pigeons

successfull}r

learned

to

search

for

the

tar-get

and

they

were

later

tested

with

display

sizes of

3,

6,

9,

and

12,

'rhe

_pigeons

in

the

[Rirget-Expansion

Group

showed

higher

search accuracy scores and shorter reaction times

than

the

_pigeons

in

the

Target-Contraction

Greup.

The

_pigeons

thus

demonstrated

a search asvmmetry

between

targets

as shown

in

previ-ous studies

in

humans.

We

discuss

the

results

in

relation

to

optic

flow.

Key

words; expansion; contraction; search asymmetry; optic

flow;

_pigeons

Optic

flow

is

thought

to

be

an

irnportant

cue

for

visual ani-mals, especial]y

for

those without

_good

stereoscopi[ vision, with regard

to

_perceiving

depth

of motion,

Approaching

and receding objects

_generate

expanding and contracting

images

on

the

observer's retina,

Using

retating

logarithmic

$piral

pat-terns

that

are

known

to

induce

illusory

cxpansion and

con-traction

in

humans

without changes

in

the

patterns'

physical

size,

Martinoya

and

Delius

(1990)

reported that

_pigeons

could

discriminate

an expanding spiral

from

a

contracting one.

However,

conclusive evidence

demonstrating

that

_pigeons

per-ceive spirals as either expanding or centracting

has

not

_yet

been

obtained.

The

_present

study explored

pigeons'

perception

of

expansion and centraction using avisual search

_paradigm,

Tletkeuchi

(1997)

showed

that

a search asymmetiy

in

humans

exists

between

targets;

that

is,

searching

for

an expanding stimulus among contra[ting ones

is

easier than the reverse.

Methods

Subjects.

Feur

pigeons

CColutnba

tivia)

were arbitrarily

di-*

L

Center

for

Frentier

Science,

Chiba

University,

1-33

Yhyoi-cho,Inage,Chiba263-8522,Japan

E-mail:nnakamura@chiba-u,jp

This

work was supperted

by

the

Grants-in-Aid

for

Scien-tific

Research:

21-1994,

10623465

to

N,

Nakamura.

"le

thank

Haruka

Nishimote

for

her

assistance

with

the

pres-entstudy.

CopY

vided

into

two

_groups.

Tlvo

_pigeons

in

the

Target-Expansion

Group

were required

to

search

for

an expanding

spiral

(the

target)

among contracting ones

(the

djstractors).

Two

_pigeons

in

the

TleLrget-Contractien

Group

were required

to

search

for

a

contracting spiral

(the

target)

among expanding ones

(the

tracters).

Apparatus

and

Stimuli.

We

used

four

identical

operant

conditioning chambers

C32

×

35

×

34

cm),

A

rotating

mic spirai

_pattern

(50

×

50 pixels,

]OO

pixels=29.7

mm) was

used

for

both

target

and

distractor

stimuli.

In

humans,

when

the spiral

is

rotated

in

the

same

direction

as

it

is

wound,

it

is

perceived

as a contracting

_pattern.

When

the

spiral

is

rotated

in

the

opposite

dLrectien

as

it

is

wound.

it

is

perceived

as an

expanding

_pattern.

Either

left-

or right-wound spirals could

appear

in

each visual

search

displaB

in

order

te

prevent

the

geons

from

responding

based

on

difft]rences

in

spiral winding

directions

between

target

and

distractors.

No

distra[tor

ed

in

phase.

The

spiral rotated at

54001s.

Each

spiral appeared

in

one ef

the

16 _possible

invisible

cells

(60

×

60 _pixels)

in

an

imaginary4

×

4squarematrix.

Procedure.

Pretraining,

After

a

S-s

intertriaL

interval

(ITI),

a

black

cross appeared

in

the

center of a white

display

area.

A

_peck

at

the

cross extinguished

it.

After

O.5

s,one

target

and

5 distractors

appeared

in

the

dispLay

area;

that

is,

the

play

size was

6. A

siiigle

_peck

at

the

target

extinguished _the

search

display

and was reinforced

by

3-s

access

tu

mixed

grains

with a

flash

of the

hopper

light,

A

_peck

atany

distractor

(2)

The Japanese Psychonomic Society

NII-Electronic Library Service

TheJapanesePsychonomic Society

88 The

_Japanese

_Journal

of

Psychonomic

Science

Vol.

31,

did

not extinguish

the

search

display

and

did

not

Iead

to

a

timeout.

Each

trial

continucd until

the

_pigeens

_pecked

at a

tar-get,

A

correct search respense was

defined

as

the

first

_peck

di-rected

onl}r

at

the

target

and not atany

distractors.

A

randotn-b]ock

_procedure

directed

the

selection of winding

directions

and

target

locations

among

64 trials,

so

that

all

_possib[e

target

locations

at which

_pigeons

_pecked

were reinforced equally

by

food.

This

continued unti]

the

_pigeons

attained at

least

a

6e%

accuracy rate

during

asingle session.

1'raining.

This

phase

was

identi[al

to

the

Pretraining

phase,

except

for

the

fo11owing

_points.

A

_peck

atany

distractor

extinguished

the

search

display

and resulted

in

a

3-s

timeout,

fo11owed

by

a correction

_procedure,

These

cerrection

trials

were not

taken

into

account

for

trial

counts or

_performance

scores.

The

number ef

trials

_per

session was

increased

to

256 trials

by

introducing

an

intermittent

reinforcement

_procedure.

The

_probabi]ity

of

_primary

reinforcement

(fbod)

was

de-creased

to

02S.

The

_pigeons

were

trained

ibr

a minimum of

seven sessions ef

256 trials

each, and the accuracv criterien

'

was

defined

as

90

96

or

_greater

correct

respenses

during

a

sin-gle

session,

The

_pigeons

_in

the

Target-Expansion

Group

reached thjs criterion,

However,

the

_pigeons

in

the

Target-Contractjon

_Group

_failed

to

reach

it

even aftera maximum of

21

sessjens.

As

the

birds

in

this

latter

_group

demonstrated

no

difference

in

mean accuracv scoies

between

the

first

seven

'

training

sessions and

the

final

seven

training

sessions,

the}r

also advanced

to

the

Tlesting

_phase,

described

below,

ksting,

The

pigeons

were

tested

with

display

sizes of3,

6,

9,

and

12,

A

random-block

_procedure

directed

the

selectien of

windi ng

directions,

target

lecations,

and

display

sizes,

Each

pi-geon

completed

24 testing

sessions.

Reaction

time

(RT)

in

each

trial

was

defined

as the time

between

the

appearance of

the

search

display

and the subject's response

to

_(i,e.,

a

_peck

at)

one of the spirals

in

the

display

(the

target

or a

distractor).

Median

RTs

for

correct responses

(RTs<IUOms

and

R'I's>

2750

ms were eliminated) were calcuLated separately

for

each

display

size within a

t"'o-session

block,

and

the

calculated

Rts

for

each subject were averaged over

the

24

sessions.

Results

Pretraining,

The

pigeons

learned

the

discrimination

quicklv

't'he

number of sessions required to reach

the

60%

criterion were two

for

asingle

_pigeon

in

the

Thrget-Expansion

Group

and

five

for

the

other

three

_pigeons.

TVaining.

The

mean accuracy scores

in

the

fina]

seven

ses-No,1

sions were

93.9%

in

the

'i'arget-Eipansion

Group,

and

74.3%

in

the

1'arget-Contraction

Group,

respectively,

A

one-way

ANON'LA

with

_group

(1larget-Expansion,

Target-Contraction)

as

the

between-subject

variable revealed

that

the

main effect ef

groupwassignificant(F(1,2)=37.08,p=.026),

1lrsting.

"rhereas

the accuracy scores

for

the

farget-Con-tractien

Group

tended to

decrease

as a

function

of

display

size,

the

scores

for

the

Ilarget-Expansion

Group

remained

high

across

display

sizes.

A

twe-way

ANOVA

with

group

as the

be-tween-subject

variable and

display

size as

the

"Fithin-subject variable revea[ed significant main efTkcts of

_group

(F(1,

2)=

33.06,

_p=,029)

and

display

size

(F(3,

6)

==-61,84,

p<.OOI),

as

well as asignificant

interaction

(F(3,

6)=25,56,

_p<.OOI).

The

RTls

in

the

'l'arget-Expansion

Group

were sharter

than

those

in

the

Target-Contraction

Group,

A

two-way

ANOX{A

revealed a significant main effect of

display

size

(F(3,

6)

z'

12,62,

_p=,O05)

and a marginally significant main effect of

_group

(F(1,

2)="

18.46,p--･-,OSO),

Discussion

Our

pigeons

shewed a search asymmetry

between

expan-sion

and contraction

targets.

Searching

for

an expanding spiral

among contracting ones was easier than the reverse,

_lf

the

pi-geons

respended

based

en

differences

in

rotation

direction

be-tween

target

and

distractor

spirals,

this

search asymmetry should not

have

occurred.

The

search asymmetry was

likely

due

te

two

_possible

factors:

a

2D

factor

and

a

3D

factor.

The

2D

factor

was

the

diference

in

apparent size

between

target

and

distractor

spirals;

this

difference

was examined

in

the

present

study

The

3D

t'actor

was

the

difference

in

sensitivity

to

approaching and receding stimuli,

In

many cases, an ob-server

is

required

to

detect

approaching objects more rapidly

than

receding objects.

The

_present

study

did

not examine

whether

the

_pigeons

_per[eived

depth

based

on

the

rnotion of

the

rotating spirals.

Further

research expioring

the

effect of the

3D

factoi

on the search asymmetry

found

in

the

_present

study

will

lead

to a more extensive

discussion

of optic

flow

in

pi-geons,

References

Martinoya,

C.,

&

Delius,

J.

D. (1990).

Perception

of rotating

spiral

_patterns

by

_pigeons.

Biologicat

(lybernetics,

63,

134.Takeuchi,

T,

_(1997),

Visual

search of expansion and