P-1A-6 並立FIRIスケジュールにおける弁別刺激変化の効果(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

TheJapanese.Psychonomic Society

zaa

koanese

.fournat

of

1tsv'cltonomic

SLienvc

2VOZ,Vol.Zl,Nol1,37 3S

Prizewinner's

SummaryP-IA-6

Effects

ofdiscriminative

stimuli

under

concurrent

and

random-interval

schedules

fixed-interval

Taku

Ismi

and

Kleio

Takayuki

Uitiversily'

SAKAGAMI

In

a

discrete-trial

_procedure

we examined

the

choice

behavior

of

pigeons

under concurrent

fixed-interval

(FI)

30-s

and

random-interval

3e-s

schedules.

In

one condition, a

discriminative

stimulus

that

was associated with

the

FI

schedule remained unchanged

through

_the

_trials.

In

the

other conditions,

it

was changed

after

a

fixed

period

(9,

15,

or

21

s)

had

passed

since

the

onset of

the

coneurrent schedules.

In

the

latter

conditions,

the

pigeons'

responses

to

the

FI

scheclule were

suppressed

in

the

periocl

before

the

discriminative

stimulus was changed.

This

result suggests

that

in

the

present

procedure

the

immediacy

of a reinforcer

had

a

minimal

effect

on

choice,

'

Key

words

:

fixed-interval

schedule,

discriminative

stimulus, choice,

key

_pecking,

_pigeon

The

present

experiment

exarnined

the

choice

of

'

pigeons

under

concurrent

fixed-interval

(FI)

and

'

'

random-interval

(RI)

schedules.

An

FI

schedule

provides

a

reinforcer

following

a

first

response after a

fixed

waiting

period,

An

RI

schedule, which

is

analo-gous

to

a variabre-interva]

(VI)

schedule,

probabilis-tically

_provides

a reinforcer at

any

m6ment

in

the

waiting

period.

Nevin

(1971)

showed

that

in

concur-rent

FI

and

VI

schedules

a

pigeon's

relative rate of response

_to

an

FI

schedule

gradually

increased

as

the

waiting

_period

approached

its

end.

This

_pattern

of

response would

be

cuntrelled either

by

the

immediacy

of

a

reinforcer

or

by

time

as a

discriminative

stimu-lus.

The

immediacy

of a reinforcer, which

is

a recip-rocal of

the

delay

to

a

reinforcer,

is

known

to

be

one

of

the

major

determinants

of

choice.

The

sooner a schedule

_provides

a reinforcer

_the

more

likely

it

is

to

be

chosen.

On

the

other

hand

a

discriminative

aspect

of

time

has

been

acknowledged

in

studies with a

single

FI

schedu]e.

Under

an.

FI

schedule

the

response

of an animal

is

withdrawn

in

an

early

_part

of

the

waiting

_period,

and more and more responses

are

made

as

the

waiting

_period

approaches

its

end.

In

other

words,

animals

discriminate

the

time

when a

'

_Departrnent

_of

_Psychology,

_Keio

_University,

2-15

45

Mita

Minato-ku,

Tokyo,

108-8345

reinforcer

is

]ikely

to

be

provided.

In

thc

_present

experiment we examined

the

extent

to

which

the

immediacy

of a reinforcer, and a

discriminative

aspect of

time

affected

choice,

by

manipulating

a

stimurus

that

signalled

the

waiting

_period.

Method

Subjects

Five

heming

pigeons

were maintainecl at

about

8096

of

their

free-feeding

weights.

They

had

experienced a variety of choice experiments,

Apparatus

The

experiment used

three

stanclard

operant

chambers,

each

with

three

response

keys

on

the

front

wall.

The

keys

were

arranged

horizontally

and

each

of

them

could

be

transilluminated

by

either

a red,

green,

blue,

yellow,

or white

light,

The

aper-ture

of a

food

hopper

which vLras

located

below

the

center

key

was

i]luminated

by

a

white

light

when

it

provided

access

to

mixed

_grain,

A

houseiight

on

the

rear wall

_provided

_general

irlumination.

Procedure

A

discrete-trials

proeedure

was used

and

an experimental session consisted of

60

trials

(one

triar

in

every

45

s).

1.

Normal

condition.

At

the

beginning'of

a

trial,

the

center

key

_was

transilluminated

with

white

light.

A

_pecking

response

on

the

key

turned

off

the

white

'

light

and

turned

on

the

lights

of

the

left

and

the

right

'

The Japanese Psychonomic Society

NII-Electronic Library Service

TheJapanesePsychonomic Society

38

The

Japanese

Journal

green

key

was

associated

with an

FI

schedule, and

the

red

key

was

associated

with

an

RI

schedule.

The

FI

and

RI

keys

appeared on

the

left

and

right

keys

randomly and equally often.

The

FI

schedule

pro-vided6s of access

to

food

fo]]owing

a

first

response

3es

after

the

onset

of

the

two

schedules.

The

food

availabirity

of

the

RI

schedule was controlled

by

a

probability

gate

of

O.0333

tested

by

_pulses

at

1s

intervals

so

that

the

nominal

interfood

interval

was

30s.

When

food

from

the

RI

schedule

became

avail-able a

first

response

_produced

2s

of access

to

the

food.

The

_pigeons

could

therefore

maximize

the

arnount of

food

in

a session

if

they

always chose

the

FI

schedule.

When

food

was

presented

from

one

of

the

schedules

the

availability

of

food

from

the

other

schedule was cancelled and all

illumination,

except

the

hopper

light,

was

dimmed.

After

the

food

presen-tation

an

intertrial

interval

began

in

which

the

only

illumination

was

the

houselight,

2.

Conditions

of

either

9,

15,

er

21s.

In

these

conditions

the

FI

key

was

yellow

for

either

the

first

9,

15,

or

21

s,of

the

30-s

waiting

period

of

the

FI

sched-ule.

The

key

was

blue

for

the

remainder

of

the

'

period.

All

of

the

other

_procedural

details

were

the

'

same

as

those

in

the

normal

condition.

All

of

the

conditions

were

conducted

for

one session

per

day

during

a

two

week

period.

The

order

of

the

experimental conditions was as

follows:

for

two

pigeons,

Nonmal,

15

s,

9s,

and

21

s; and

for

the

other

three

pigeons,

15

s,

Norrnal,

21

s, and

9s.

Preceding

each

of

the

choice conditions noted

above

the

pigeons

experienced

training

sessions.

A

training

session was

the

same

type

of

session

as

the

next choice condition except

that

only one of

the

two

schedules

was

_presented

in

a

trial.

Each

ef

the

FI

and

the

RI

schedules were

_presented

30

times

in

a

training

session

in

a

randomized

order.

Results

and

Discussion

We

divided

the

30-s

waiting

_period

into

ten

3-sbins,

and

calculated

a relative rate of response on

the

FI

key

for

each

bin,

for

the

last

5

sessions of each condition,

for

each

pigeon.

The

relative rates of

the

responses averaged over

the

five

_pigeons

are

shown

in

Figure

1.

In

the

Normal

condition,

the

relattve rate

of

PsychonomicScience

Vol.

21,

No.

1

1,O

go.

g

o.s

NE

.!

g

o.6

rd

a

E

:

o.4

geg

ts

g

o,2

ke

o,e

123456789le

Bins

Figure

1.

The

average relative rates of responses

on

the

FI

key

for

the

five

_pigeons.

The

numbers

of

the

bins

for

which

the

relative rate was

lated

are･

indicated

en

the

abscissa.

The

broken

lines

indicate

a

change

of

the

stimulus

of

the

FI

key.

increased

gradually

as

the

30-s

waiting

period

ap-proached

its

end.

In

the

other

conditions,

responses on

the

FI

key

were suppressed when

the

FI

key

was

yellow.

The

responses

increased

gradually

after

the

FI

key

became

blue.

In

addition

the

maximal relative

rates

were

higher

in

these

conditions

than

in

the

N6rmal

condition.

The

immediacy

of a reinforcer

does

not explain

these

results

because

we

kept

the

reinforcement schedules constant

across

all

of

the

conditions,

Therefore

the

_gradual

increase

of

the

relative

rates

of

the

responses

in

the

Normal

condi-tion

ean

be

interpreted

as a result of

temporal

dis-crimination,

In

the

9-,

15-,

and

21-s

conditions,

the

change

in

the

color of

the

FI

key

seemed

to

serve as a

time

murker

that

signalled

when

to

commence

timing

the

interval

before

food

presentation.

When

a

time

marker

eccurrecl

earlier

the

curve

(Figure

1)

of

the

relative rate after

the

time

rnarker was

Shallower.

This

indicates

that

as

the

length

of

the

intervar

which

was

to

be

timed

increased,

the

response

rate

in

that

interval

increased

rnore

_gradually.

This

result

agrees

with

_past

research on

timing.

References

Nevin,

J.

A.

1971

Rates

and

_patterns

of responding

with

concurrent

fixed-interval

and variable-interval

reinforcement.

Ibumal

of

the

Empen'menlal