Differential effects of spatial separation on visual feature binding in humans and pigeons(Summary of Awarded Presentation at the 29th Annual Meeting)

The Japanese Psychonomic Society

NII-Electronic Library Service The JapanesePsychonomic Society

The1inPanesgfoumol of As)'chonomde'Sciencg

2011,VnL 30,Ne.1,145-146

Summary

ofAwarded

Presentation2-345

Differential

effects

of

spatial

separation

on

visual

feature

binding

in

humans

and

_pigeonsi)

Sho

OTAKI*･

*2,

Kazuhiro

GoTo*,

and

Shigeru

WATANABE*3

K),otoUbeiversity*,

_fopan

Society

_for

thePV'omotion

of

Science*2,

Keio Uhiversicy*a

We

examined themechanisms of visual

feature

binding

inhumans and pigeons, two visually

dorninant vertebrates thatuse differentneural substrates forfeatureprocessing. InExperiment 1,

we examined

how

humans and _pigeons

bind

color and lineorientation.

Subjects

were trained to

search

for

a target among

distractors

consisting of

horizontal

and vertical

lines.

In

the

feature

condition, one of the target lineswas a differentcolor than the

distractors.

In

the

binding

condition, the target and distractorswere characterized by differentcombinations of colors and orientations.

Both

species located

the

target

fastcrunder thefeaturecondition compared with the

binding

condition, suggesting that

binding

requires

focal

attention inboth species. InExperirnent

2,we examined the effects of spatial separation on

visual

feature

binding,

The target and

distractors

consisted of

two

horizontallyaligned colored Iines.The distance between thetwo

lines

had

little

effect on target

localization

in

humans,

whereas

the

_process was

hampered

in

_pigeons.

These results may refiect

differences

in

theneural substrates of these species.

Key

words:

feature

binding,

visual search, pigeons

Pigeons and humans are visuaLly dominant

verte-brates. Previous studies of the two species have

shown

functional

similarities

in

the perception of

shapes, colors, and pictures. The two species, how-ever, showed functional differences in other

processes,such as _perception of partly occluded

fi-gures and hierarchicalcompounding stimuli, These

results suggest that pigeons are

less

capable of

inte-grating

features

in different

dimensions,

including

colors and shapes, compared with humans.

In

the

present study, we examined the mechanisms of

vis-ual

feature

binding inhumans and

_pigeons.

General

Methods

Subjects.

This study included six adult humans and three homing _pigeons.

Apparatus.

All

stimuli were _presented on

18.l-* _{Graduate School of Letters,Kyoto University,}

Yoshida-hommachi,

Sakyo-ku,

Kyoto

606-8501,

JapanL}This

work was supported

in

_part by

Grants-in-Aid

for

Scientific

Research

{22.7370

and 22700271) and the

Global

Center

of

Excellence Program

(D029)

from the Ministry

of

Education,

Culture,

Sports,

Science

and

Technology

of

Japan.

Copyright2011

inch or 10.1-inch,touch-sensitive monitors

for

the

humans and _pigeons, respectively.

Procedure.

Subjects

searched

for

a single target item among a tota]of 3,7,11,or 15,items.The target

and distractorsconsisted of a set of two different

feature

dimensions,

such as color

(red

or _green)and

lineorientation

<horizontal

and vertical), The target

differed

from

the

distractors

in

either one of thetwo

featuredimensions

{feature

condition) or both of the

feature

dimensions

(binding

condition}.

Correct

re-sponses resulred ina chime sound forthe humans or

a food reward for the _pigeons, whereas incorrcct

responses resulted

in

a

buzzer

souncl

fer

humans

or a

15-stimeout

for

the _pigeons.

Experiment

1

Inthe

first

experiment, we examined how humans

and _pigeons

bind

color and

line

orientation.

The

target contained red horizontal and _green vertical

lines.For thefeature condition,

distractors

consisted of two lines,one of which was the

_.tt

same color

(red

or

green) as the targetirrespectiveol orientation. The other

line

was another color

(blue).

For

the binding

condition,

distractors

contained two lines

that

were

the same colors and orientations as thetargetbut the

,

The

Japanese

Psychonomic Society.Allrights reserved.

The Japanese Psychonomic Society

NII-Electronic Library Service The JapanesePsychonomic Society

146 The

_Japanese

_Journal

of Psychonomic

Science

Vol.

30,

No. 1

(a)T15oogt3oom.gMoe-$

900=B

7.

£

sco

B'

d'Humans(b)tooo

950

900850

seo

750

700

Pigeons

37

11

15

37

3f

t5

Displaysize

Figure

1,

Response

times

for

correct trials

under the binding

(circles)

and feature

conditions

(squares)

in Experiment

1

for

humans

(a)

and _pigeons

(b>.

Schematic

representations of search displays forthe two

conditions are shown together.

combination of the

features

was switched.

Results.

Figure

1shows themean correct response

times

for

each condition and species. Both species

located the

target

fasterin

the

feature

condition

compared with the

binding

condition, Response

times were independent of display size under the

featurecondition, whereas _they

increased

as

display

size increased under the binding condition,

Two-way repeated measures analysis of variance

{ANOVA)

revealed significant main effects for the

condition

[humans:

F{1,5)= 109.83,

p<

.Ol;

_pigeons:F

(1,

14)=

139.18,P<,Ol]

and

djsplay

size

[humans:

F{3,

15)=58,43,

_P<,Ol;

pigeons:

F(3,

42)=20,91,

_p<,Ol].

The interaction

between

these _parameters was also

significant

[humans:

F(3,

15)=42.70.P<.O1;

_pigeons:

F(3,42)=4.69,P<.Ol].

Experiment

2

We

then examined how spatial separation affects

feature

binding.

The

targetand

distractors

consisted

of two vertically aligned linesthatwere colored as in

Experiment

1,

Two

lines

were adjacent or separated.

Results.

Figure 2 shows the mean correct

re-sponse times

ior

each condition and species.

Separat-ing

thelineshad littleeffect on target

localization

in

humans

but

hampered

it

in

pigeons.

In

both

species,

three-way repeated measures ANOVA revealed

sig-nificant main effects

for

the condition

[humans:

F(1,

5)=28.00,

_P<.Ol;

_pigeons: F(1,14)=214.62.

_p<,Ol]

and

display

size

[humans:

F(3,15)=:22.03,

_p<.Ol;

pigeons:

F(3,42)=22,03,

_P<.Ol].

-

There

was aiso a

-17oo 17eo

g

tSOO

t500

ggoe

goe

g7ee

7oe

esoo

soo

37 lt 15

37

tl 15 Disptaysize

Figure2. Response

times

for

correct trials

under the

four

conditions

in

Experirnent

2

for

humans

(a)

and _pigeons

_(b).

ADJ, SPR,

BDG,

and FTR

indicate

adjacent, separate,

binding, and feature conditions, respectively.

significant interaction between _the condition and

display

size

for

both

species

[humans:

F(3,15}=

14.48,

_P<.Ol;

_pigeons: F(3,42)=22.66,

_p<,Ol].

Al-though

no other main effects or

interactions

were

significant inhumans, an additional significant main effect of spatial separation

[F(1,

14)=96.49,

_P<.Ol]

was observed

for

pigeons.

Discussion

The results oi Experiment 1suggest thathumans and _pigeons share common mechanisms

for

binding

stimulus color and orientation, which requires focal

attention. The results of Experiment 2,on the other

hand, suggest thatthespecies

differ

in

theirabilities

to

bind

spatially separated

features.

Features such as colors and orientation are _processed

in

cortical regions of the primate

brain

(Livingstone

&

Hubel, 1988) and the midbrain of

birds

(Wang,

Jiang,

&

Frost,1993).

We

thereforehypothesize that

feature

binding

also occurs at a relatively

later

stage

in

primates compared with avian brains. The

diffe-rences

in

neural substrates

in

these two species may

underlie the differentialeffects of spatial separation

on

feature

binding.

References

Livingstone, M.

&

Hubel, D.

_<1988).

Segregation

of

form,color, movement, and depth;anatomy, physi-ology, and _perception.

Science,

240,740-749.

Wang,

Y.

C.,

Jiang,

S.

&

Frost,

B,

J.

(1993).

Visual

processing in pigeons nucleus rotundus: Lumi-nance, color, motion, and loorning subdivisions.