Global and local visual processing by chimpanzees (Pan troglodytes)(Summary of Awarded Presentation at the 25th Annual Meeting)

The Japanese Psychonomic Society

NII-Electronic Library Service

The JapanesePsychonomic Society

7}zelapanesefournat o.f PsychenomicScience

20e7,VoL 25,No.2,281-282

Summary

ofAwarded

Presentation2P30

Global

and

local

visual

_processing

by

chimpanzees

(Pan

troglodytes)

Toyomi

MATsuNo*･

** and

Masaki

ToMoNAGA*

Primate

Research

institute,

Kyoto

Universily,*

JLiPan

Societ),

for

the

Promotion

_of

Science**

Two

experiments

investigated

_the visual processing of hierarchica]stimuli

by

chimpanzees,

Using a delayed matching-to-sample

{DMTS)

task, fivechimpanzees were testedfortheir

tual

dominance

ofgLobalor localinformation inhierarchicallydesigned figureswhich consisted of

globalconfigurations and localelements.

The

results of thefirstexperiment

displayed

individual

difference$inglobal-localprocessingef thefigures.Two of

the

chlmpanzees exhibited a advantage

forlocallevelprocessingand theothers exhibited an advantage

for

_processing

_globalshape.

The

results of the second experiment rcvealed _that manipulation of the denstty of localelements

affected the _performance of thethreechimpanzees who had _{globalprecedence} but did not affect

the performance of theether two chimpanzees who

had

local

precedence. These result$ suggest

thata precedence for_global or localvisual _processing

in

chimpanzee$ isnot

determined

a priori

but depends on strategies which are learned individuallyto solve a given task,

Key words: hierarchicalstimulL _{global-]ocalprocessing,}chimpanzees

Human vision _{tends to process}

_global

configur-ation

before

_proceeding totheanalysis of local

fea-tures

(Navon,

1977).On thc other hand,non-humans,

such as

birds

and monkeys are

known

to possess a

clear advantage

for

the processing of localvisual

elements rather than

_global

shape, Previous reports

on chimpanzees, the closest relatives of

humans,

are

controversiaL Fagot and

Tomonaga

₍₁₉₉₉₎

com-pared the visual search performances for

hierarchi-cal stimuli of chimpanzees and humans, and

found

that chimpanzees did not show any clear tendency

for_{global precedence.}

In

a study

by

Hopkins

and

Washburn

_{2002)

using aDMTS procedure,however,

chirnpanzees showed an advantage for _processing

global configuration of stimuli. The _purpose of this

study isto vary several _parameters, such as the

stimulus size and

density

of

local

e]ements, which

would be infiuentialto global-local processing of

hierarchica]

stimuli, and reconsider the _problem.

Method

Fivc chimpanzees

{5-28

_years

in

age) _participated

inthe experiments.

One

of them had experience in

* _{Primatc Research In$titute,}

_Kyoto

_University,

Kanrin. Inuyama,

Aichi

484-8506

This research was supported by MEXT _grants

{l6002001,

16300084) and

_JSPS

Research

lowship(16/1060).

Copyrtght2007

participating

in

a visual search study using

hierar-chical stimuli

CFagot

& Tomonaga, 1999),The

stim-uli were displayedon a 22-inchCRT monitor

equip-ped with a touch-screen.

The stimuli were hierarchical _geometric figures

(Figure

IA)

subtending visuai angles of

approxi-mately 21,7,10.8,5.4,or 2.7degrees

in

height

and

width of _globalconfiguration at a viewing

distance

of 45cm.

The

size of the localelements was one-eleventh ofthe size of theglobal configuration.

Each trialof a DMTS _task was

initiated

when the

subject touched a warning stimulus

located

at,the bottom of the screen. A sample stimulus was pre-sented

for

500

m$ atthecentre of thescreen after the

response and disappearance of the warning stimulus.

Aftera de]ay_periodof 500 ms, two alternatives were

presented

A

subject was required totouch the same

figureas thesample.

During

the testingthree types of trialswere

pre-sented. The type of trialdepended on an {ncorrect alternative which differedfrom _thecorrect

alterna-tiveineither

its

globalconfiguration

_("global"

condi-tion), its

local

elements

("local")

or both ef these

("beth").

Priortothe testsessions

the

chimpanzees were trained with trials of the "both"

condition to

attain at criterion

for

learning

which was set as

>85% accuracy,

In

the

first

experiment, the

chim-panzees were tested in the 3 condition$ and the

The Japanese Psychonomic Society

NII-Electronic Library Service

The JapanesePsychonomic Society

282

The

Japanese

Journal

of Psychonomic

Science

VoL

25,

No,

2

A,Stimuli

if

:TT v --v + v-v +vvvvvy-- -v- VVWWVw v- v+ -- vy v- vv v -wwvW V V B.Results

loo Ayumu loo

g

gO

cy

vD"`(X'9

ggo

t80 t so o o e o

sI:

,,svx.,,

geI:

WVvv w----v -V-WW v

:

WVWyvvvv Yvw Cleo YY

::v-v-:

:

::

vv vY'Y +both-global -local

"""`' 50 50 2,7 5,4 10.8 21,7 2.7 5,4 10,82a.7 stimulus size stimulus size

Figure 1. Examples of stimuli

(A)

and results

(B)

inExperiment 1. The graph on the

left

displays the results of the subject who used

local

cues, and thegraph on the right

depicts

the results of the subject who

demonstrated

aglobal precedence.

stimulus size was varied.

In

the second experiment

the separation between the

local

elements was

dou-bled

by

decreasing

the number of elements.

Results

Inthe firstexperiment, response tendenciesof the

chimpanzees were

inconsistent

between subjects

(Figure

IB).

Two of them displayed a considerable

advantage

for

_processing

local

elements: they wcrc

highly successfu}

for

the "both"

{87.1%

_correct

in

average} and "local"

conditions

(87,1%),

with all of

thestimulus sizes.

Their

_performance was _poor how-ever forthe

"global"

condition

(62.7%).

These

results

suggest that they focused on the ]oca]elements of

the

hierarchical

stimuli and neglected the _global

con-figurations.Incontrast, theother chimpanzees

per-formed

the trialsof the"globaV' condition

(85,9%)

as well as those of the "both"

condition

{89.2%),

but

made many errors of discriminatieninthe shape of

localelemcnts of the"local" condition

(61.296).

These

results suggcst thatthey attended tothe global

con-figuration

of the$timuli and that the

local

elements spilled out of theirperceptualrepresentation.

Inthe second experiment, the _performance of the

two chimpanzees who

demonstrated

local

prece-dence

was littledlfferent

from

that

in

Expcriment

1,

On

the other

hand,

the performance of thesubjects

who

disp]ayed

aglobalprecedence was influencedby

the decreased density of the localelements.

This

occured _particularly when the stimulus size was

larger

("global"

conditien with the largeststirnulus

size: 80.6% correct

in

Expcriment

1,

and

64,6%

in

Experiment 2).

Discussion

The

presentstudy clearly

dcmonstrated

individual

differencesofglobalor localprocessing of

hierarchi-cal stimuli by chimpanzees. Two of the five

chim-panzees

displayed

a _{globalprecedence} and

the

other

demonstrated a

lecal

precedence.

In

addition, one of our subjects who had a strong tendency of a global

precedence was the same chimpanzcc who did not

show any evidence of globalprecedence

in

thestudy

by Fagot & Tornonaga

<1999).

These difference$

could be caused by differencesof what was learned

in

the training phase when

the

subjects could use

cues of both the global conliguration and local

ele-ments to successfully discriminate thefigures,The

large

variations of stirnulus size used inthisstudy

could

_also

_promote such individual differencesin

strategies tosolve the

discrimination

task, To sum-marize, a precedence

for

a global or

local

visual

proc-essing by thechimpanzees which was demonstrated

in

this

kind

of

behavioral

experiment couLd not

be

determined a _priori,

but

was

largely

dependent

on what each individual

learned

according to thetask

and stimulus configurations used inthe task,

References

Fagot,

J.

&

Tomonaga,

M

1999

Global

and local

processing inhumans

(Hbmo

saPiens) and

zecs

{Ilan

trogtodytes):

Use

of a visual search task

with compound stimuli.

joumal

of

ComParative

Rs))cholQgy,113,3-12.

Hopkins,

W,

D.

&

Washburn,

D,A. 2002 Matching

visual stimuli on the basisof _globaland local

tures

by

chimpanzees

(llan

trogZod),tes)and rhesus

monkeys

(ildiacaca

mutatta),

Animal

CQgnition,

5,

27-31

Navon, D. 1977

Forest

before

trees:

The

precedence

of global

features

in

visual _perception.

Cognitive