P-1A-11 ラットにおける遅延による報酬の価値割引報酬量効果の検討(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

The Japanese?sychonomic Society

V'he

_JaPanese

_Jot{rnat

_of

lli.v:honomic

･Stience

2e02,Vol.2I,No.1,39-40

Prizewinner's

Summary

P-IA-"11

The

effect

of

reward

amount

on

discounting

of

delayed

rewards

in

ratsi)

Daisuke

SAEKi,

Tetsuo

YAMAGucHi,

and

Masato

ITo

Osaka

Cits,

Uitiversib'#

The

present

study used a

concurrent-chains

schedule

to

examine

the

effect

of

reward

amount

on

delay

discounting

in

rats.

Delay

_periods

dcfincd

by

fixed-interval

schedules

in

the

terminal

links

were

fixed

at

2

s

in

one alternative and varied

in

the

other alternative across

five

conditions ranging

from

2s

to

40

s.

The

reward amounts were varied across

three

conditions

by

manipulating

the

number

of

food

_pelrets

(1

pellet

vs.

1

pellet,

5

pellets

vs.

1

pelret,

and

5

pellets

vs.

5

pellets).

The

discounting

rate

estimated

with

a

modified

hyperbolic

function

was

higher

in

the

smaller reward

amount condition

than

in

the

]arger

reward amount conditions.

This

finding

indicates

that

in

animals

the

rate of

delay

discounting

decreases

us

the

reward

arnount

increases.

'

'

Key

words : choice,

clelay

discounting,

reward amount effect, concurrent-chains schedule, rats

Delayed

rewards

are

judged

rower

in

value

than

immediate

rewards

because

the

delay

is

thought

to

cliscount

the

value of a

reward.

Studies

of

delay

discountjng

in

human$

reported

that

the

discounting

is

well

described

by

the

folrowing

hyperbolic

func-tion:

v-i+Ant)

a)

where

V

is

the

discountecl

value

of

the

delayed

reward,

A

is

the

reward

amount,

D

is

the

length

of

the

delay,

and

k

is

a

free

_parameter

representing

the

cliscounting

rate, respectively.

On

the

other

hand,

some

modifications

of

Equation

1

have

been

proposed

in

studies

of

delay

discounting

in

animals.

The

following

modification,

a

hyperbolic

function

with an exponent, was

proposed

by

Ro-driguez

&

Logue

(1988)

because

rnany studies of

animar choice assume no

_proportional

(or

linear)

relatiOn

between

the

reward amount

(A)

and

the

subjective value

(

V)

:

'

_Department

of

Psychology,

Faculty

of

Literature,

Osaka

City

University,

3-3--138

Sugirnoto

Sumiyoshi,

Osaka,

558J8585

1)

The

first

author

was supported

by

Research

Fellowships

of

the

Japan

Society

for

the

tion

of

Science

for

Young

Scientists.

As

V ±

1+to

(2)･

where s

is

a

free

parameter

representing

the

sensitiv-ity

to

A.

Some

studies

of

delay

discounting

in

humans

repor-ted

that

the

cliscounting

rate

decreases

as

the

reward

amount

increases

(e,g.,

Green,

Myerson,

&

McFad-den,

1997),

however,

this

effect

has

not

been

reported

in

animal

studies.

Therefore,

the

present

study used rats

to

examine

the

effect of

the

reward amount on

the

discounting

raEe

in

terms

of

Equation

2.

Method

Subjects

Five

experimentally

naive

male

Wistar

rats

were maintained at approximately

80%

of

their

free-feeding

weight.

Apparatus

Two

standard experfmental chambers with

two

horizontally

arranged response

levers

"iere used.

A

45-mg

food

_pellet

was

used

as

a

reinforcer.

Procedure

The

subjects

were

exposed

to

a

concurrent-chains schedure

in

which a single

variable-interval

30-s

schedule arranged entry

into

one ef

the

two

terminal-links

with equal

_probability.

Fixed-interval

schedules were

in

effect

for

the

terminal

links.

The

length

of

the

de]ay

was

fixecl

at

2

s

for

one

(ress

delayed)

alternative and was varied

for

the

The Japanese Psychonomic Society

NII-Electronic Library Service

TheJapanesePsychonomic Society

40

'

The

Japanese

Journal

of

'

other

(rnore

delayed)

alternative

acruss

five

condi-tions

(2

s,

5s,

10

s,

20

s,and

40

s),

The

number of

food

pellets

presented

in

the

terminal

links

was varied across

three

conditions:

three

_pairs

of

1

_pellet

vs.

1

pellet,

5

perlets

vs.

1

_pellet,

and

5

_pellets

vs.

5

_pellets

were

arranged

for

the

more

delayed

and

the

less

clelayed

alternative, respectively.

A

time-out

_period

followed

the

shorter

delay

to

equate

a

trial

duration

between

the

two

alternatives.

AII

of

the

subjects were exposed

to

each of

the

delay

and reward arnount conditions.

Each

conditipn remained

for

at

least

14

sessions and was continued until a stability criterion

(Ito

&

Asaki,

1982)

was

satisfied.

Results

and

Discussion

Figure

1

shows

the

mean choice

proportions

of

the

groups

for

the

more

delayed

alternative as a

function

of

the

delay

for

each

reward

amount

condition.

The

choice

_proportions

"Tere

calculated

by

dividing

the

initial-link

responses

for

the

rnore

delayed

altevnative

by

the

total

initiar-link

responses,

Equation2

was

tt

fitted

to

the

choice

proportions

represented as

'

tion

3:

B,.

L

btr],

(3)

B.+B,

-bIG,.

VL

where

B

is

the

initia]-link

responses,

b

is

a

free

parameter

representing a

bias,

and

M

and

L

are

the

more

delayed

and

less

delayed

alternatives,

respec-tively,

The

sensitivity

te

the

reward

amount

(s)

can

be

specified

only

in

the

5

_pellets

vs.

1

_pellet

condition

:

the

terms

representing

_the

reward

amount

(A")

are

cancelled out

in

Equatien3

for

the

other conditions,

The

mean

le

values

for

the

five

subjects were:

O.049

for

the

1

_pellet

vs.

1

_pellet

condition;

O.026

for

the

5

pellets

vs.

1

_pellet

condition;

and

O.030

for

the

5

pelrets

vs.

5

_pellets

condition.

The

mean

fe

value

for.

the

1

_pellet

vs.

1

_pellet

condition

was

significantly

higher

than

the

values

for

the

other

two

conditions

(t(4)=2.37

for

the

5

_pellets

vs.

1

_pellet

condition and

t(4)==2.42

for

the

5

_pellets

vs.

5

_pellets

conditien,

Ps

<

,05).

However

the

difference

between

the

mean

k

PsychonomicScience

Vol.

21,

No.1

u 1.0gg'

:i:.2

e o,i=g.

E'

:/:a

}

o.4sI. n.i'6 e O.2S

ij

D.: o lo 2o 3o 4oo lo 2o 3o 4Do lo 2o 3e 4o

Delay

(scc)

Figurel.

TAe

mean

choice

_proportions

for

the

more

delayed

alternative as a

function

of

the

delay

for

three

_pairs

of reward amount.

The

curves

are

the

best

fit

for

the

modified

hyperbolic

functions

(Equation

3).

1peUeLvs.1pelletSpclEetsvs.1pellet5peUetsvs,5pellets

k

±

D.044b.IA3rL=O,97

k..O.v]6s=o.gab-O,71ff\O.98

k.O.024b=1.32ri-,D2

values

for

the

5

_pellets

vs.

1

perlet

condition

and

the

5

pellets

vs.

5

_pellets

condition

was not

significant.

These

results

indicate

that,

in

rats,

the

rate

of

delay

discounting

decreases

as

the

reward amount

increases.

This

finding

is

consistent with an

investiga-tion

by

Grace

(1999).

The

study examined

the

effect of reward amount on

the

sensitivity

to

de]ay

esti-mated with

_the

_generalized

matching

law.

However

the

data

did

not

yield

a statistically significant

difference

(Grace,

1999).

The

fact

that

the

direction

of

the

reward

amount

effect

is

the

sarne

for

humans

and animals suggests

that

this

effect

is

a

general

phenomenen

on

the

dis-countirig of

the

delayed

rewards.

References

Grace,

R.C.

1999

The

matching

law

and

dependent

exponential

discounting

as

accounts

of

self-control choice.

Iburnag

of

the

Etpen'meninl

Anab'sis

_of

Behavior,

71,

27-44.

Green,

L,,

Myerson,

J.,

&

McFaddefi,

E.

1997

Rate

of

temporal

cliscounting

clecreases

with

amount

of

reward.

Memory

&

Cagnition,

25,

715-723.

Ito,

M.,

&

Asaki,

K.

1982

Choice

behavior

of rats

in

a

concurrent-chains,schedule:

Amount

and

delay

of

reinforcement.

foumal

of

the

Empen'mental

Analy-sis

of

Behavior,

37,

383-392.

Rodriguez,

M.L.,

&

Logue,

A.W.

1988

Adjusting

delay

to

reinforcement:

Comparing

choice

in

pigeons

and

humans.

Iburnal

of

E}xPen'meental