Objective estimation of the state of content understanding by Near-Infrared Spectroscopy (NIRS)(<Special Edition>Brain Imaging and Psychonomic Science)

7he

_.laDanese

_foutvtalofIlsychonomic Science

2009,VoL 28,No,1,2 16

Original

Article

Objective

estimation

of

the

by

Near-Infrared

state

of

content

understanding

Spectroscopy

_(NIRS)i>

Nobutaka

ENDo,

MasayoshiNAGAI,and

Takatsune

KuMADA

IVbtional

Jnstitute

ofAdvanced

industrial

Science

and

71gchnotog),

(AJST)*

Three experiments used near-lnfrared spectroscopy

(NIRS}

toexamine brainactivation inthe

frontal cortex during comprehending and reporting thecontents of a story.

In

Experiment

1,the

participantswere asked toview humorous comic strip stories under thefollowing two conditions:

comprehending and reporting the contents after the experiment

(report

conditionL and not

reporting the contents

(non-report

condition). The changes of

Oxy-Hb

concentration

in

the left

prefrontal region were higher in the report condition than in the non-report condttion. In

Experiments

2A

and

2B

the participantswere asked toarrange sets of cards, according toeither

story developrnent or shape comp]exity, without reporting

the

content Dfthe stery.

The

changes

of

Oxy-Hb

concentration

in

the

left

and right _prefrontalregions were

higher

for

card arrangements

related to story development than for those related to shape comp]exity. When articulatory

suppression was used, activation inpart of

the

left

prefrontalregion was reduced. The anterior

portion of the

left

inferior

frontal

gyrus

(around

BA

47)

was consistently activated when

partici-pants actively triedtounderstand content, regardless of the type ot task and memory load. We

suggest a new objective method, using

NIRS,

with which tomeasure the

brain

activity associated

with trying tounderstand content.

Key

words: content unclerstanding, near-infrared spectroscopy, objective estimation

When

passing on

information,

it

i$

important

to

determine whether the information has been

cor-rectly understood,

For

example,

in

the

field

of

educa-tion, teachers can adjust the amounts and/er levels

of

information

that

are appropriate

based

on

stu-dents' understanding ifthey can cerreetly assess

each student's understanding.

Substantial

efforts

have

been

devoted

to establishing techniques

for

rneasuring how well _people understand information.

i) _{This research was supported by the} "Assistive

technologies to ensure safe and comfortable

lifesty]es

for

persons with

disabilities"

prograrn of the

SpeciaL

Coordination

Funds

for

ting Science and Technology, 2004, from the

Ministry of Education,

Culture,

Sports,

Scjence

and

Technology

(MEXT),

Japan.

*

_{Corresponding}

_should

_be

_{sent to}

_Nobutaka

higashi, Asaminami-ku, Hiroshima 731-3195

(e-mail:

[email protected]) or Masayoshi Nagai,

National Institute of Advanced Industrial

Science

and

Technology,

Central

6,

1-I-1

Higashi, Tsukuba 3e5-8566

Ce-mail:

[email protected]).

In

contrast, no method

has

been

developed

to

eva!u-ate how much effort isengaged in conducting that

task.

The

_purpose of the present study was to

sug-gestan objective rnethod toshow whether _people are making efforts tounderstand content,

Currently, two methods are avai]able tomeasure

human

comprehension. One isthe subjective report.

In

subjective reports, students or participants rate

their own understanding level

(e.g.,

rating their

com-prehension using a

five-pojnt

scale

from

"understood

complctely" to

"could

not understand at aU">.

Al-though

this

method iscommonly used ina variety of

research areas, itsmain weakness isthat

_it

isnot

objective.

For

example, some _people may report that

they understDod the rnateria] very well even when

they actually didnot understand itat all. Incontrast,

others may report that they did not understand the

material even when they really did. Itisnot

reason-ab]e to compare subjective rating scores between

individuals because criteria used in rating may

differ.

The

subjective reporting method

is

also not

reliable when applied tochildren or people with

NII-Electronic Library Service

N.

ENDo,

M.

NAGA[,

and

T.KuMADA:

Objective

Estirnation

of

Understanding

by

NIRS

3

niLive impairment, because it

is

rnore diMcult

for

them tomonitor theirown states of understanding

than

it

is

for

healthy

adults.

Anotiner

popular

method for measuring _people's comprehension is

tests of content understanding.

This

rnethod

in-volves asking people questions about thecontents of

a story or narrativ'e. This should _providc a mere

obiective result, and

it

makes

it

possible

for

test

scores tobe compared between individuals. This isa

reasonable method

by

which

to

eva]uate whether

peop]e understand the _pTovided material, However,

neither

this

method nor subjective rating can show

how much effort

is

expended toward the goal of

understanding

the

materiaL

For

example,

tn

cases

in

which subjective rating showed "complete]y

under-stood contents" and content-understanding test

showed

futl

marks,

it

is

unclear whether

the

_person

actually made any effort tounderstandi the material.

Ifthe contents were too easy, he/she might hardly

make an effort.

In

contrast,

ii

the contents were

diMcu]t

and were barely understood, helshe should

have

rnade rnore effort.

[n

other cases

in

which the

subjective rating shows "did not understand" and the

content-understandlng

test

showed a low score, it

is

also unclear which was true:the person may

haye

activel'y triedtounderstand the content, but failed

(made

more effort), or he/she may not even have

triedtounderstand atall

{made

no efforO.

Although

it

is

important foreducational purposes to determine

whether the content

has

been

correctly understood,

objectively rneasuring

how

much effort is

invested

in

understanding or to what

degree

children or

stu-dents arc trying to understand

is

also

important

in

selecting the appropriate levelof materials to

maxi-mize

learning.

We developed a new way tomeasure brain activity

related

to

trying

to understand content using

near-infrared

spectroscopy

(NIRS).

NIRS

is

an optical

method that non-invasively estimates temporal

changes

in

concentrations of

Oxy-hemoglobin

(Oxy-Hb), Deoxy-hemoglobin

(Deoxy-Hb),

and

Tota]-hemoglobin

(Total-Hb)

based on the amount of

ab-sorbed near-infrared iight,using the modified

Lam-bert-Beer

Law.

Using

the

relationship

between

neu-ral activity and regional cerebral

blood

fiow,

a

NIRS

system shows activated

brain

regions where

concen-trations of Oxy-Hb are observed

CVillringer

and

Dirnagl,

1995).

Brain

monitoring

by

NIRS

has

sev-eral merits. The equipment iscompact, making it

easy tomeasure

brain

activity with fewer physica]

limitations.

More

kinds

of tasks can

be

_performed

using NIRS than with other brain imaging

tech-niques, such as

fMRI

and

MEG.

Our main

purpose

was to specjfy brain regjons

involved

incontent understanding using

NIRS.

In

particular,

we

focused

on whether people actively

tried tounderstand something. Ifthe state of trying

'

tounderstand can

be

evaluated

by

monitoring

brain

activity using

NIRS,

we can _provide a new _paradigm

for

measuring

human

cognitive processes

for

neuro-science studies. This brain-imaging method has the

advantage of showing specific

brain

activity, which

cannot

be

determined

using subjective rating or

un-derstanding tests.This method can _provide

immedi-ate

feedback

on the

brain

aetivity required totry to

understand content, which wilL for example, allow

teachers toselect appropriate materiaLs

for

their

stu-dents or rnake

it

_possible

towarn

drivers

who have

lost

their

focus while driving.

We

used two

kinds

of task toencourage people to

understand content: testing of content

understand-ing and sequencing of story events by arranging

cards.

Although

therequirements of thesetwo tasks

are essentially different,_partjcipants inboth tasks

are required to actively try tounderstand the

con-tent, Ifthe specific

brain

activation associated with

the process of understanding can

be

consistently

observed regardless of

differences

in

thesetasks,this activation will be a usefu] index to evaluate _people's states of understanding.

Experiment

1

In

Experirnent

1,

we used a comic strip $tory to

induce

brain

activity related totryingtounderstand. We

focused

on the anterior prefrontalarea torecord

brainmeasurements,

because

pilot experiments

indi-cated a steady signal inthisarea.

Methods

Ftzzrticipants,

Thirteen

healthy

young

Japanese

4 The

_Japanese

_Journal

of Psychonomic

Science

VoL

28,

No.

1

adults

(mean

age=22.4 years, SD=1.40, nine rnales,

four females) who were nai'vc to the _purpose of the

study _participatedas _paid volunteers.

All

of

partici-pants were right-handed and had normal or

correct-ed-to-normal vision. Handeclness was determined

using the

Japanese

translated version of the

Edin-burgh Handedness Inventory

_(Oldfield,

1971).

Writ-ten

informed

consent was obtajned

from

all partici-pants.

Stimuli

and tasle.

Six

short

hurnorous

comic strip stories were used as stimulL Each story conslsted'of a series ef four

frames.

Four of the six stories

in-cluded original and complete content, the logical

development of which involved an introduction,

de-velopment, turn,and conclusion

(i.e.,

standard story

deve]opment). Four ofthe series were equal interms

of theirunderstandability and

leve]

of

humor,

based

on the results of priorratings by nine

participants

who did not _participate inthisexperimenL In

con-trast,the remaining two series consisted of

four

frames that were randomly chosen from different

stories and consequently made no sense

(i.e.,

scram-bled

stories).

Each

frame

of thc story was _prcsented

for 5 s inthe center oi a 22 inch cathede ray

tube

(CRT)

monitor

(Iiyama

MA203DD),

so that the total

duration

ef each story was 20s. Participants

ob-served the comic strip stories under the following

three conditions.

In

the

first

condition

(the

report

condition), the normal stories were presented on the

CRT

monitor and _participants were

instructed

to

understand why the stories were funny. After the

experiment, allof participantswere required _togive

a titleto the stories and report the contents and

funn}r elements. In the second condition

(the

non-report condition), the normal stories were _presented

and _participants were

insi.ructed

to

observe the

sto-ries

but

were not

instructed

toreport on their

under-standing of them.

In

the thirdeondition

(the

scram-bled

conditionL participants were

instructed

to ob-serve the scrambled steries, which could not be

un-derstood.

Two of the

four

normal stories were

ran-domly assigned to the repert cendition and the other

two to the non-report condition.

Each

condition was

repeatcd twice, and the order of the conditions was

counterbalanced across _{participants.} Each trial

be-Experjment 1 Pre-time. Stimutatfon :10sec 20 sec lesec Rest30 sec 30 sec Experments 2A and 2B

Pre-time

Stimulation 60 sec

ilOsec 30 sec

10 sec 40 seC _loo'sec

Figure

1,

Experimental

schedule.

gan with a 1O-srest before the stimulation, followed

by

thecemic strip presentation

for

20

s.

Each

presen-tationwas followed by a 30-srest. The experimental

schedule

is

shown

in

Figure

1,

Recording,

We

used a multichannel near-[nfrared

specLroscopy

<NIRS)

system

(ETG-4eOO,

Hitachi

Medical

Corp.)

to

measure

brain

activity.

Thel

NIRS

unit was operated at near-infrared wavelengths of

695 and 830 nm and was used

to

measure tem,poral

changes

in

concentrations of

Oxy-Hb,

Deoxy-Hb,

and

Total-Hb.

We

used a _pair of _pro'beholders, each

consisting of a

4

×4array.

Sixteen

optodes

(eight

emitting and eight detccting) constituted 24

chan-nels ineach _probe. The distance

between

each

emit-tingand thecorresponding

detecting

optode was

3.0

cm. The sarnpling rate foreach channel was apprexi-mately 10Hz.

Probes were

placed

on the

tronta]

region of each

hemisphere

(Figure

2).

Cz

in

the

international

10/20

system was set to align with the connecting

line

between the _positionsat thesuperiQrlposterior

cor-ners of the arrays

(i.e.,

the emitter

for

the right probe

and the detector for the leftprobe), and Fz was

aligned with the middle of the two connecting lines

between the _positions at the superiorfanterior

cor-ners of

the

arrays and

between

the

_positionsat

supe-rior/seeond

from

anterior

intersections

of the arrays.

Because particlpants'heads were not exactly the

sarne size, the

Fz

posjtion

differed

slightly among

participants, The position of each column of emitters

or

detectors

in

the

arrays was _paralle]to

the

midline of the

_bTain.

_We

examined brain activation in

Chan-ne]s

1-7

in

both

hemispheres as the region of

in-tercst,according toresults

from

a pilot experiment.

Data analysis. All ef thc tria]swith movement

Rest60

NII-Electronic Library Service

N

ENDo,

M,

NAGAi,

and T.KuMADA:

Objective

Estimation of Understanding

by

NIRS 5

E･I

Emitter Detector

Channel

Figure 2. Pesitions of optodes _p!aced

in

line

with the prefrontal regions

in

both

hemi-spheres.

The

distance

between

each emitter

(indicated

by light gray squares) and the

corresponding detector

(indicated

by dark

gray squares) was

3cm,

These emitters and

detectors

were placed

in

probes with

4

×

4

arrays, NIRS responses at

Channels

1-7

in

both hcmtspheres were analyzed in

ment

1,

whereas those at all channels were

analyzed

in

Experiments

2A

and

2B.

Nasion anterior

wao-@ew@-@@@@

"@me.@m(Dma

@@@@

ss@"@ww@-@@@@

-@ma@m@wa

Fz,:

cz='

ewo-@ma@"

CD@@@

m@wa@m@wa

@@@@

me(i})M@ee@-@@@@

-@ge@-ge

:/s

H

posterior 3cm wam Inion

o

artifacts, determined

by

steep changes in the time

series of the concentration of

both

Oxy-I-Ib

and

De-oxy-Hb, were removed from the analysis, The raw

data

of Oxy-Hb and Deoxy-Hb from individual

chan-nels were normalized,

because

the raw

NIRS

data

were originally relative values and could not be

aver-aged

directly

across participants or channels.

The

z-scores forthe normalized data could be averaged,

regardless of the unit of measurernent

{Matsuda

and

Hjraki, 2006; Shimada and Hiraki, 2006;

Otsuka,

Nakato, Kanazawa, Yamaguchi, Watanabe, and

Kakigi,

2007).

The

g-scores

for

changes

in

Oxy-

and

Deoxy-Hb during the task

_(z-Hbt,.i,}

were calculated

as follows:

z-Hbta.k=

HbtaskJMean(Hbp,..t,.,)

.

SD(Hbpre-time)

Hbpre.tiine

and

Hbtask

represent

Hb

data

in

the pre-time

and the task _periods, respectively.

A

mean and a

Standard

deviation

for

changes

in

Oxy-

and

Deoxy-Hb

in

the totalpre-time period were ca!culated

for

each trjalin each channel, and the z-score o'feach

data _point

in

Oxy-

and

Deoxy-Hb

during

the task

period

in

each trial

in

each channel was calculated

based

on that trial'sand channel's mcan and

stan-dard

deviation,

Prevlous

studies indicated that the

hemodynamic

responses typically

lagged

behind

the

stimulation by a few seconds. AdditionaUy, some

task-related

brain

activity should occur after the end

of the comic strip _{presentation,} Especially

in

the

report conditien,

it

was likelythat participant$

ac-tivelytriedto remember the content of the story

after the _{presentation.} Thus, we defined the comic

strip presentation

(20s)

as the stimulation period;

both the stimulation and the rest

(30

s) _periods wcre

defined

as the task _period; the last

10s

of the re$t

periodwas

definedi

as the _{pre-taskperiod}

for

the next

presentation,

In

preparation for

N'iRS

data

analysis,

thee-scores of

Oxy-

and

Deoxy-Hb

were

filtered

with

therunning average over 5.0sand calculated at each

point

in

the time series

data,

The

averaged z-scores

of Oxy- and

Deoxy-Hb

were calculated foreach

chan-nel and each condition for individual _{participants.}

The

data

were compared among the task conditions

using repeated-measures analysis of variance

{ANOVA),

using the

Greenhouse-Geisser

correction.

These data calculation and statistical _procedures

were also used toanalyze the results of

Experiment

2,

Results

and

Discussion

Al]_participantscould reasonably _provide a titleto

the cornic strip stories _presented

in

the

report

condi-tion,and they could a]so correctly recal] the

con-tents,

Accordingly,

we confirmed that they

corn-pletely understood why the stories were considered

humorous. A total

_gf

five trials with movement

artifacts were excluded

from

the analysis

(1

trial

in

the report condition, 1 in the scrarnbled condition, and 3

in

thenon-report condi,tion).

Overall,

Oxy-Hb

concentrations inthe report

con-ditiontended toincrease during and after thecomic

presentation

in

the

left

and right prefrontalregions

(Channels

1and 4

i,n

the lefthemisphere and

3

and

7

in

theright

hemisphere;

Figure

3),

However,

no such

trend was observed

in

the other two conditions.

In

con-6

The

Japanese

JournaL

of

Psychonornic

Science

VoL

28,No.

1

A

Ch 1

8sm=Aas

UEgbt:tit TEme[se[] Ch4 ChS Ch2 Ch] 10 20 30 40 50 Time[se[1 5 o

-560

O Ch6

li

ll

'

i'"

/

+t-tttttntL-"t1'

i /10 20 30 40 SO 60 Time[sec] Ch - Report uapha-- Non-report

iiii""i",-"t

Scrembled so

-so

lo 2a 3o 4o se 6o Time[sec] 7 5e-so s o

-5

la 2o 3o 4a so so o Tlme[se[]

tt

tl

_/1

tt

tt

tt

tt

: :

..,..,,..,,L..1",,

tl

!/

'10

70 30 4D SO 60 Time[se[]

:-so

//

_//

//

//

//

_//

//

/

t-t

i'

//

//

t/lo

2o 3o 4o so 6a Time[se[1 Ch1

Bg

:T E:o

f:t

.5o

Time[secl Cb4 ChS 5 S; :'

ll

l: D O :-h

...l,....

I/ i'

-5

-S

a lo 2o 3o 4a so 6o o lo 2o 3o Tlme [sec]

Figure

3.

Time

course of

the

averaged

periods

in

the

(A)

left

and

(B)

right

indicatcd between the vertical dash

Ch2 la Clt3

-5

2D ]a 4o so so o lo 2e 3D 4o so 6o Time[secl Tlme[sec] Ch6 Ch7 511 S 11 11 11 o. o 1 "'.-".es F' 11

-

-s

6o o lo 2o 3o 4o so eo o lo 2o ]o 4o so 6e Tlrne [sec] Time [sec]

in

Oxy-Hb concentration during the _pre-task and

Comic

strip stories werc _presented during the

40 50 Tlme[sec] change

hemispheres.

ed lines. tasktirne

ditiontended todecrease during the task _period in

the

lcft

and right prefrontal regions

(Channels

1

and

4 in the lefthcmisphere and 3 and 7 inthe right

hemisphere;

Figure

4).

To

demonstrate

the

Lenden-cies of the actixration during each of thestlrnulation

and task periods,we analyzed

data

for

each period

separately.

The mean z-score$ of

Oxy-

and Deoxy-Hb

concen-trations during the stimulatiQn

for

each channcl

were subjected te two-way ANOVA with channel

and taskcondition as the main factors,For Oxy-Hb

concentratjons inthe lefthemisphere, the analysis

showed a significant interaction between channel

and task condition:

FC2.19,

26.29}=5.26,P<,Ol).

In

the right hemisphere, none of the main effects or

possib]e interactions was significant. The channels

activated inthe report condition compared with the

other conditions are shown inTable 1. For

Deoxy-Hb concentrations, none of the rnain effects or

possi-ble

interactions

was significant ineither hemisphere.

We also assessed the Hb concentration changes

during

the

task

period. For

Oxy-Hb

concentrations

in the lefthemisphere, the analysis showed a

sig-nificant

interaction

between

channel and task

condi-tion:

F(2.07,

24.87)=7.45,P<.O05.

Inthe right

hemi-sphcre, a significant interactionwas found between channe! and task condition:

F{1,71,

20.49)=3.96,

_P<

.05.

The channels activated

in

the report condition

compared tothe other conditions were

found

in

both

hemisphcres

(Table

1). For

Deoxy-Hb

concentra-tions,none of the main effects or possib]e

interac-tions was significant ineither hemisphere,

The

left

and right prefrontal regions were

acti-vated when _participantszctively tried t.ounderstand

the cornic $trip stories and to memorize their

acti-NII-Electronic Library Service

N.

ENDo,

M,

NAGAi,

and

T.KuMADA:

Objectjve

Estimation

of

Understanding

by

NIRS

7

A

Ch 1

89eGv

ts2usSg Tlme[sec] Ch4 Ch2 ChS 10 2D 30 40 50 60 Tlme[sec] Tlme[sec] Tlme[sec] Ch3 Ch6Time[secl t Report ww Non-repert "i"i"-"i" Scrambled lo 2e 3o 4e so so Tlme [se[] Ch7 Time[sec]

B

so Ch 1 8?eeu tsi8ss Time[sec] Ch4 Ch5 Ch2 10 20 30 40 SD EO Timelsec]Ch6 Ch3 TLme[sec]Ch7 5D seii'ii'/

=

-S

-5

-5

-S

o lo 2o 3o 4o so so o la 2o 3o 4o so 6o o lo 2o 3o 4o soso o lo 2o 3o 4o so 6o

Time[set] Time[se[] Tlme[sec] fime[se[1

Figure

4, Time course of the averaged change

in

Deoxy-Hb

concentration

during

thepre-task ancl task

periods

in

the

(A)

Ieftand

(B)

right

hemispheres.

Comic

strip stories were presented during the time

indicated

between

the vertical

dashed

Iines,

vation of theIeft_prefrontalregion was important for

understanding the stories,

because

thisregion was

activated during the stimulation. This activation

was not obscrved

in

the

non-report or scrambled

conditions, and no difference

between

thetwo

condi-tionswas observed. Oxy-Hb concentration changes

were

likeLy

more sensitive

to

the understanding

process

than were Deoxy-Hb concentration changes, and

Deoxy-Hb

changes

had

less

variation

during

the

task period than did

Oxy-Hb

changes. Thus, itseems

possible to evaluate whether _participants tried to

understand thecontents of the comic strip stories

'by

monitoring Oxy-Hb concentration changes,

How-ever, itj,sa]so likelythat our results were caused

by

interactionsamong the _processes of understanding,

memory, anclfor motivation. Because _participants

had

to

actively remember the contents

in

order to

report them after the experirnent, the memory load

could increaseandfor themotivation might bemuch

higher

in

the report condition.

We

addressed this

issueinExperiment 2A.

Experiment

2A

InExperiment 1,we showed that activation inthe

left

prefrontal region was useful

in

eyaluating

whether a _person was trying tounderstand content.

In

Experiment

2A,

we exarnined thispossibility

us-ing

another task thatdidnot require an active

mem-ory _process, but that controlled ior the motivation

factor.

The task inExperiments 2A and 2B was to

sort _picture cards insequential order. Experiments

2A

and

2B

were

identical

except

for

thetask required

of _{participants.}

Methods

1]kerticipants.Fifteen healthy _young

Japanese

8 The

_Japanese

_Journa]

of Psychonomic

Table

1

Channels

activated

in

thereport condition,

compared with thenon-report and the

scrambled conditions

in

Experiment

1.

Science

VoL

28,

No.

1

Stimulation

period

Task

period

Mean eh Condition k-score Mean ch Condition z-score

Leftlreport

1.30

non-report

-

1.72* scrambled

-2,20***

4report

2.03

non-report

-

l.87* scrarnbled

-2.18***

l

report 1,62 non-report

-

1.46** scrambLed

-

1.72**

4report

2.71

non-report

-

1.48** scrambled

-1.48*'

Right

3report O.62

non-report

-1.76*

serambled

-1.l3

7

report

O.59

non-report

-

!.79*

.s.

cram,bled

-1.2.5.

*P<,05, *"EP<.Ol, **#EP<.eOl: _{NewmanJKeuls test}

for cemparison with mean a-scores in the report

eondition.

adults

<mean

age == 22.7_years,

SD

= 1.80,10_males, five

females)

who were nai've tothepurpose of the study

participated as _paid vo]unteers.

All

_participants

were right-handed and had normal or

corrected-to-normal vision.

Handedness

was determined using

the

Japanese

trans]atedi version of the Edinburgh

IIandedness Inventory

_(O]dfielcl,

1971). Written

in-formed

consent was obtained

from

all_{participants.}

Stimuli and task, We used two types of _picture

cards as stirnuli

in

Experiments

2A

and

2B.

One

type

consisted of sets of _picturecards used inthe _picture

arrangement testof the

Wechsler

Adult

Intelligence

Sca]e

(WAIS),

Unlike the simp]e observation of the comi ¢ strip story in Experiment 1,inExperiments

2A

and

2B,

_participants were asked

to

arrange

the

picturecards sequentially to make reasonable stories

(the

story condition>. The other type of stimuLus

consisted of sets of cards printed with nove],

rnean-ingless shapes

_(Endo,

SaikL

Nakao,

&

Saito,

2003).

Participants

were required toarrange the cards

se-quentially according tothe subjcctive complcxity of

shapes

(the

shape condition). Although _participants

needed toarrange the cards

by

hand

in

both

condi-tions,in the story condition, _participants were

re-quired to actively understand the content of each

card, as well as

the

context and reasonable

develop-ment of the given story.

However,

unlike

Experi-ment 1,_participants did not have to remember the

contents of the story after the task period.

In

con-trast,in the shape condition, only _perceptual

judg-ments of shape complexity were required.

Each

task

condition was repeated threettmes,and participants

perforrned the two tasks aLternately. The orcter of

the

conditions was counterba]anced across

partici-pants. Eaeh trialbegan with a 10-srest, fellowed by

card sorting

'for

30

s,and then a

60-s

rest

{Figure

],).

In the story condition, participants were instructed

to

actively understand

the

story content and

deter-mine the reasonabLe order of thestory within 30s. In

the shape condition, _participants were instructed to

determine

the

reasonable order of cornplexity

in

the

sets of novel shapes using a consistent criterion. If

participants eempleted the sorting before the time

limit,

another set of _pictures

(or

shapes) was _given.

After sorting the cards,

participants

were asked to

relax

during

the rest period.

Recording.

The

NIRS

s}rstem, the measured chan-nels, and the _placement of the optodcs were identical

to those

in

Experjment

1,

whereas the tasks and

rnaterials _presented were different. The channels

examined in Experiment 1 were selected based on

the results of the pi]otexperiment with the task of

Experiment

'L.

but

it

was not clear whether the

se-lectedchanneis would

be

appropriate tothe task in

Experiment

2,

Thus,

we needed toanalyze a wider range of brainregions, and we examined all channels

{i.e.,

24

channeis

in

each

hemisphere>

because

this

task was likelytoresult

in

brain activities in

differ-ent regions than those monitored inExperiment l.

Data anatysis,

We

focused on

the

Oxy-Hb

concen-trationas _the index of

brain

activation

because

in

Experiment

1,

Oxy-Hb

was rnore $ensitive to the

understanding process

during

the

task period than

was Deoxy-Hb. In

fact,

in

many

NIRS

studies,

changes

in

Oxy-Hb

concentration are primarily used

inanalyses

(e.g.,

Matsuda & Hiraki,2006;

Otsuka

et

al,,2007) because they are most sensitive tochanges

in

regional cerebral

blood

fiow

and are correlated

sig-NII-Electronic Library Service

N. ENDo, M,

NAGAi,

and T.KuiaADA:

Objective

Esttmatton

of Understanding

by

NIRS

nal

(Hoshi,

Kobayashi, & Tamura, 2001; Strangman,

Culver,

Thompson,

&

Boas,

2002>.

All trtalswith movement artifacts were exc]uded

from

analysis, and

the

Oxy-Hb concentratien raw

data were norma]ized as z-scores. We analyzed the

data regardless of the accuracy of the card arrange-ment

because

we

focused

on

brain

activation that

was related totheattempt tounderstand, rather than

tounderstanding something correcdy.

Results

and Discussion

One

trial

in

the shape condition was excluded

from

the analysis because of movement artifacts. In the

story condition,

the

Oxy-Hb

concentrations

tended

Ch1 S20 20i

o

t

[.. I

::

10 IOI i:o oi

/Lt

.

).

t

6-lo

.lo

:･ O !O 40 60 SO iOO O 10 Ch4 Tlme[se[] ChS 20 20 10 10 o o

-10

-10

O ZO 4TU/rnefi[Psec]60100chsO 20 4TOimee[Osec]80 2ai i 2oi il i 10: ID: ll i ",: o,

-lo,

ii

,

-rn,

i

/t

/

O 20 40 fiO BO 100 O 20 Chll T/me [sec] [h12 20 20 IO 10 o o

.10

-10

o 2o 4o Go eo loo o 2o 4o 6o so 20 ZOi ]/ 10 101 i e o l

.LO

-10i

20/1 20 10Ii 10 _o i _a :1

-le

_,

:/

_,

-10

o' 2o

ko

6o sD ioe o 2o 4o 6o eo Tinie[sec] Ch22 Jjme[sec]

t/ 2o: / ?n :1 10:1 10 :1 ol o

-10,

i 1'

,

-10

O 20 40 SO 80 100 O 20 Time[se[]

Figure 5.

Time

course oi the averaged change

Experiment

2A.

Card

sorting was performed

lines.

9

to increaseduring and after card sorting inthe same

Ieft

ancl right _prefrontalregions

(Figures

5

and

6)

as

in

the report condition inExperiment 1:Channels 1

and

4

in

the

]efthemisphere and

3

and

7

in

the

right

hemisphere.

To examine the trends

in

activation

during each stimulus

(during

card sorting) and task

<durjng

card sorting and resting) periods,we

]yzed thedata foreach _period separately. The mean

z-scores of

the

Oxy-Hb

concentrations

during

card

sorting foreach channel were subjected totwo-way

ANOVA

with channel and task condition as main

effects.

In

the

left

hemisphere

(Figure

5>,theanalysis

shewed significant main effects o'[ channel: F(3.30,

46.23}=9.11,

P<.OOI;

and

task

condjtion:

F(1.00,

ChZ [h] i 20:1 l

'

ii 10: 1

---

Starv :/ :l

.iOai

I

nv

shnpe 40 io uo lop o 2D 4D 6e ao 100

fime[seel ChE Ime[EE[] Ch7

2o 2oi1 :1 10 tO:1 11 o o, 11

-10

-101

1

//

10e o !o 4o Sv Bo mO o 20 40 60 SO loD Chg T/me [set] ChlO ImE [sec]

i lo:1 i :: 10 : : ::

I

.r,li

'40

6D BO tOO O 20 40 60 80 100 Time[sec] Ch13 Time [sec] Cht4

2011 ZO :1 1-,, 10 :1 o o :[

-10

･･

-10

tt

loo o 2o 4o 6e so loo o 2o 4o eo sz loD

Cht6 Time [sec] Cht7 T/me [sec]

20 il to o li

_-10

40 EO SO 100 O 20 40 fiO 80 100 Tlme[se[] Ch20 -me[sec] Ch21 20 20 10 10

o m

-10

-10

loo o 2D 4o so so leo a zo 4o 6o so mo Ut13 Tirne[sec] Ch24 T/me[yec]

20 : 1 10 : 1 :1 o _:t

-10

: : 40 60 EM tMO O 20 40 SO EO 10D TImeLsec] TimeLsec]

in Oxv-Hb concentration in the

during the time

indicated

betweenleit

hemisphere tn

10

2010o-10

2o10o-IC D

o

The

Japanese

Journal

of

Psychonomic

Science

VoL 28,No. 1

Chl ChZ Ch3

um

t/

tt

av 20:1 20 20],

at,

,

[t'

t/

''

e:mii

io ioii

-story

U,

di

um

,

:/

TiiO/, O n/ l

nv

shape

it :1 1:

o-la: /

-10

-lo/

/

Chg Jlme[Eec] ChS -me[se[] ChG Tir/e[se[1 Ch]

t/

ZO:1 20 20 :1 tO:: 10 10 :t D, u o :/

-10ii

-10

.tm

?o 4o Eo Bo lon o ?n 4n 6e eo too o ze 4e Ee so loo o 2o 4o Eo so loo

Time[se[l th8 _"me[secl Chg _T/me[sit] ChlO _-me[secl

//

ZOII 20 20 11 10;/ 10 10 :: O, D O 1:

-10

･

･

-lo

-te

o' 20

to

Eo ao mo o 2o

to

6o so MO O 10 "o 60 8e loO

Chll Time [secl Ch12 1'i,"E[se[1 Ch13 T/me [sec] Ch14

2o 1[ 20ii :: lo lo loiI _{O O} l/ D

,

11

-10

-10

-1011

//

2o 4o Eo SO 100 O 20 4o SO So 100 O ]O 4o so eo 100 O 20 40 fiDeO leO

TIrae[sec] [h lS _T/me[sEcl Ch tS _I]me[sec1 Ch17 _{T/me [set]}

20 20:/ 20 il lo Iail do o ol

',

e

-10

-10,

l[

,

-10

O 20 40 eO SO 100 O 20 40 60 6o !Oo o 2o 4o 60 em loO

ChlS Ti/ne[gec] [hlg T/m"[sec] Ch20 T/melie[1 Ch21

i,

?Oil 20 2D

/t

t

/t

i 10ii 10 lo

t

/t

t

/t

i oli o o

/

tt

i

-IMil

-10

.10

20 40 60 SO 100 O 20 40 60 EO 100 O 20 4D 60 80 1,OO O 20 40 60 SO tOO

-me[sEc] Ch22 1'/mE[Ee[1 Ch23 Time[sec] Ch14 _T/me[sec]

lei i 20 20 10: : 10 10 L: 2010o.10iI io o,

-10

: o

Figure

6.

Tiine

course of the

Experiment

2A,

Card

sorting lines.

14.00)=11.25,P<.O05.

A

significant

interaction

was

also found between these factors:F(3.38, 47.38)=

4.26,

p<.Ol,

In

the right

herrijsphere

(Figure

6),

the

analysis showed significant main effects of channel:

F(2.95, 41.31)=3.69,

P<,05;

and task condition:

F(1.00,

14.00)=9.48,

_p<.Ol;

as well as a significant

interaction

between

these

factors:

F(3.49, 48,84)r-3,19,P<.05.

We also assessed Oxy-Hb concentrations during

the

task _period. IntheIcft

hemisphere

(Figure

5),

the

analysis showed significant main effects of channel:

F(2.45,

34.23}';12,79,

P<.OOI;

and task condition:

F(1.00,14.00)=

11.08,

P<

.O05;

as well as a significant

interactien between these factors:FC2,84, 39.78)=

O D

i

-tO

-10

2o 4D 6o sc too o 2m 4o En 8D tDO O 20 40 SO eO

T/me[sec] TLme[sec] Time [set]

averaged change

in

Oxy-Hb

concentration

was _perEormed during the tirneindicated

100

in

thebetwecnright

the

hemispherc

in

vertical dashed

4.39,p<.Ol. Inthe right hemisphere

(Figure

6),the

analysis showed significant main effects of channcl:

F(2.39,33.38)=6.07,

_P<.O05;

and task condition:

F(1.00,

14,OO)=-9,48,

p<.Ql;

as well as a significant

interaction between these factors:F(2.48, 34.69)･.=

3.45,

_P<.05.

The channels activated in the story

condition compared to the shape condition are

shown inTable 2.

Consistent with the results of Experiment 1,

acti-vation occurred intheprefrontalregions

during

card

sorting.

Further,

more

inferior

and _posteriorareas of

the _prefrontal regions were actlvated during and

after card sorting inboth hernispberes. Oxy-IIb

NII-Electronic Library Service

N.ENJ)o,M. NAGAi, and T.KuMADA:

Objective

EstirnatiQnef Understanding by NIRS 11

Table 2

Channels activated inthestory condition,

compared with the shape condition in

Experiment

2A,

Stimulation

period

Task

period

chDifferencechDifference

Left123456811

3.30***3,99***2.85***3.40***2.25***2.49***1.80***1.65*

124568113.44***3.87***3,83***2.61***2.58*2,56***2.28***

Right1236710

3.83**3.90**3,14**3.94**2.07**2,47**

1236710 3.33**3.45*4,02**3,85**3,48**3,04**

"P<.05,

*SP<.Ol, **"P<.OOIr

_Newman-Keuls

testforthe differenceInmean z-scores between

thestory and theshape conditions.

higher

than were

those

in

the

shape condition.

Thus,

itseerns unrikety that the activation in the anterior

areas of the prefrontalregions

(i.e,,

Channels

1

and

4

in

the left

hernisphere)

in

Experiment

1

was caused

only by memory _processes. However, the resu]ts of

Experiment

2A

could

have

been

caused

by

other

processes,such as verbalization of thescenes and the

relation between thescenes, because according tothe

study of

OkamQto,

Dan,

Sakamoto,

Takeo,

Shimizu,

Kohno,

Oda,

Isobe,

Suzuki, Kohyama, and Dan

{2004),

theactivated area corresponding to Channels 1and 4

in

the

_left

hemisphere cou]d overlap with or

sur-round Broca's area, which

is

invo]ved

in

language

processing, speech _production, and comprehension,

Infact,most of _participants reported that they used

inner

speech as a strategy

during

the picture

ar-rangement. Thus, it

is

possible thatthe brain

activ-ity

re]ated toactive processes of comprehension ob-served

in

Experiment 2A could reflect

inner

speech-related activity. We addressed thisissuein

Experi-ment 2B toidentifyspecific channel _positionsrelated

to attempting to understand content.

Experiment

2B

In

Experiment

2B,

we examined whether thc acti-vation of the left_prefrontal region

(i.e.,

Channels

1 and 4)

during

picture arrangernent was caused

by

participants' strategics, such as

inner

speech.

Our

aim was

to

specify the channel _positionsthat can be

used to assess participants'states of _tryingto under-stand content, rather than

inner

speech,

Per

se.

Methods

thrticipants. Fifteen healthy young

Japanese

adu]ts

(mean

age=22.6 years,

SD-1.20,

10

males,

five

females)

who were nai'vc to the _purpose of the study _participatedas _paid volunteers. Allof

partici-pants were right-handed alld

had

normal or

cor-rected-to-normal vision. Handedness was

deter-mined using the

Japanese

translatedversion of the

Edinburgh Handedness Inventory

(Oldfield,

1971).

Written informed consent was obtained from all

par-ticipants.

StimutL

task,and

Procedure.

We used the same

pictureand shape cards and experimental

design

and

schedule as in Experiment 2A. Participants'tasks

were also

Identical

tothose

in

Experiment

2A,

except

thatparticipants were required toperform

articula-tory suppression

(repeatedly

verbalizing thesyllable "da")

during

card sorting to

block

inner

speech.

We

rneasured the same channels as

in

Experiment 2A

and focused on changes in

Oxy-Hb

concentration,

Results and Discussi,on

No trials with movement artifacts occurred

in

ei-thercondition, and

data

for

alltrialswere analyzed.

Inthestory condition,

Oxy-Hb

concentrations in

the

left

and right _prefrontalregions tended to

increase

during

and after card sorting

(Figures

7and

8),

The mean

Oxy-Hb

concentration g-scores during the

stimulus period for each channel were subjected to

two-way ANOVA with channel and taskcondition as

main effects. In the Iefthernisphere

(Figure

7),the analysis showed a significant main effect of channel:

F(3,22,

45.03}

=;4.63,

p<.Ol;

_{and a significant}

interac-tion

between

channel and task condition: F(4,93,

68.96)=2,72,P<,05.

Inthe right hemisphere

_(Figure

12

2S10o･10

2eIUo-10

The

_Japanese

_Journal

of

Psychonomic

Science

Vol.28, No. 1

Chl [h2 Chl

9..

eglo

1[ _lo

-storv

t:t'O

O O "-.ma Shape

it

a-10

-10

.lo

O 20 40 60 8o loo o 20 40 60 SO 100 O 2o aa se BD LOO

th4 T/me[!ec] ChS Imetsec] [h6 T/me[sec] Ch7

tt

t/

?n ]o:/ 20:/

t/

t/

t/

t/

t/

t/

tt

t/

o o', ol' :1 lt

-10

-101]

-IDIi

t

/

20 4Q EO SO IOO O !O 40 EO 80. ±OO O IO 40 6D SM 10D O 20 40 EO 8D 100

T/me[se

£

] ChS Time[sec] Chg T/ffte[se1 Chln _"rne[se[]

20 20 10ii

lo 10 10il

o o oi

ii

-10

-10

-10:1

o 2o 4o 6o so ioo o 2c 4o

Ee

eo ioo o' 2o 4a Gn Ets nD

Chll T/mu [secl Cbl!

'1'ime

_[sec] Ch13 -mE [sec] [h14

20 ?Oii 201i

t/

t/

/t

IS 10 :, l 10I :,

tt

tt

t/

tt

o Di, ei'

//

/t

_.lo

//

_-lol

/t

l

.lol

l

20 -･O 6v EO IDD o ;D 40 SO eO 100 O 20 4o SO Eo Mo O 20 4o 6a He luO Ime rse[] Ch lE T/me[sec] Ch16 Tgme[se[1 Ch17 T/me[ser]

20 20 20

10 IU tD

-10

-10

-10

O ?O 40 60 80 100 O 20 4o Eo sz lom m ?O 40 eO EO 100

ClitS T/me [sec] Chlg 11m: Esec] Ch20 T/me [sec] Ch21

l PO 20 ?o

l

1 ID 10 10

i

.r,

.r,

.:,

20 qO Go Eo 10o o 2e 4m SO eo loo o 2D 4o se so loo a ZO 40 eO eo 100

Timelse[] Chl2 -me [sec] Ch23 Time [sec] [h24 T/me tse[]

t/

?O ?O 2ll:1 ii o o 2u10o.10Iil:io n

-10

O 20

Figure

7.

Time

coursc Df the

Experiment

2B,

Card

sorting

lines,

8),the analysis showed significant main effects of

channel:

F(3.30,

46,19)=3.26,

_P<,05;

and task

condi-tEon:

F(1,OO,

14.00)==6.90.

_p<.05;

as wel] as a

sig-nificant

interaction

between thcse factors:F(4.31,

60.35)=3.52,p<.oll

We alse assessed Oxy-Hb concentrations during

the task period.Inthe]efthemisphere

(Figure

7),

this

analysis showed sjgnificant main effects of channel:

F(3.27,45.74)=9.41,

_P<.OOI;

and task condition:

F(1.00,

14,OO}==

11.13,P<.O05;

as well as a significant

interactien between these factors:F{3.65,

51,15)--4.02,p<.Ol. In theright hernisphere

(Figure

8),

the

analysis showed significant main effects of channel:

F(3,42, 47.82)=8.37,

_P<.OOI;

and task condition:

o o,,

1:

-10

-10:/

4e 6o Eo 1[o o 20 40 GO SO 100 O 20 4D 6u fiO lao Tllne[se[] TlmeL'!ec] Ime[se[]

averaged change

in

Oxy-IIb

concentration

in

the

was performed

during

the time

jndicated

betweenleftthehemisphere

in

vertical dashcd

F{1,OO,

14,OO)=21.15,P<.OO1;

and a significant

inter-action between these factors:F(3.90, 54.54)=6.93,

p<.OOI.

The channels activated inthe story

condi-tio'n,compared wjth the shape condttion, are shown

inTable 3.

The areas inthe left_prerrontal region that were

activated

during

and after card sorting were reduced

when _parLicipants were asked toperform

articula-tory suppression

during

card sorting. Inparticu}ar,

the brain activation in

Channel

4 that was

consis-tently observed in Experiments 1 and 2A

disap-peared, suggesting that activation

in

Channel

4

in

the left prefrontal region w･as likely caused by

strat-NII-Electronic Library Service

N,

ENDo,

M.

NAGAi,

and

T.

KuMADA:

Objective

Estimation

or

Understanding

by NIRS

chl Ch2

_.Ch]

m ::

cn

EYIo lo lol/

-SteFy

u8 1l

E:,o o oi/ ww shape

/-

t/

S'

-to

-to

-toli

O 20 40 SO SO 100 O 10 4M Ea so loo D ?o 4o so sO.100

Ch4 T/me[!e{1 ChS 1'jme[secl ChS

.

T/me[sec] Ch7

2eii lo 20 20 10:1 10 10 10 :1 o, o o o

-10,,

ii

.10

-10

-10

O !O 4e GO 80 100 O 2o 4o eu Em mD o !o 4a eo eo

.loo

o 10 4a SO BO 100

T/me[set] ChS Tjme[sec] Chg T/merse[] ChlM Time[sec]

20 20 20

10 10 10

o o o

-.LO

.10

-10

o 2o 4o 6o so ioe o 2D 4D fio so loo o 2o 4p so so too Chll Ime [secl Ch 12 -me [sec] Chi] Tlme [sec] Ch14

2D 20 IO 20 tD 10 10 10 O U O D II le 10 1･1 10 /1 o ol' o ii

ChAe Time [sec] Chlg T/me [secl Ch20 nme [secl Ch21

20 20 201i 20i]

tt

tt

/t

//

o M Oi l, Oli

-!O

-10

-LO:1

-10:1

t/

t/

ii

i[II io io o, i

-to

: o

Figure

8.

Time

course of the

Experiment 2B. Card sorting

lines.

'

egy. In contra$t, the activation

in

Channel

1

in

the

left_prefrontal region was observed conststently in

al]experiments, regard]ess of

the

type

of task,the

presented stimulL memory

demand,

or motivation,

This suggests that a _{participants'}state of trying to

understand content can

be

evaluated

based

on this

specific

braM

activation. Incontrast, brain

activa-tion

in

the right prefrontal region showed trends

similar to those in

Expertment

2A.

but

this

activa-tion was not observed Experiment 1, A possible

interpretation

isthat the

tictivity

inthe right

hemi-sphere may reflect task-spccificfactors related to

story

formation

by

sortlng picture cards.

o o

I i

-10

-10

2z 4o 6o Eo loD O ZO 40 GO BO laO o 10 4D Eo so

Tirne[sec] Ime[sec] Tlme[sec]

averaged change in

Oxy-Hb

concentration

was _performed

durjng

the time lndicated

100 in thebetweenright the

13

hemisphere

in vertical dashed

General

Discussion

We

bave

provided evidence that trying to

under-stand content leads to activation

in

the _prefrontal

region of the lefthemispheTe

(Experiment

1)and of

both

hemispheres

(Experiments

2A

and

2B).

Changes

in

Oxy-Hb

concentrations tnthe activated

areas were consistently

higher

in

tasksthatrequired

participants]active effort tocomprehend than inthe

other tasks,

Our aim was to _previde a new _paradigm forthe

measurement of human cognitive activity related to

the

process of

trying

tounderstand content

based

on

Chan-14

The

Japanese

Journa]

of

Psychonomic

Scionce

VoL

28,

No.

1

Table

3

Channels

activated inthe story condition,

compared with the shape condition in

Experiment

2B.

Stimulation _period Task _period

chDifferencechDifference

cal _processing of words actjvated

BA

44f45,

and

semantic _processing of words activated BA 47/45

(e.g,

Poldrack,

Wagner,

Prull,

Desmend, Glover, &

Gabrieli,1999). The articulatory suppression

proce-dure

in

our study might only disrupt processing

in

Left12

3.19**5.06**

128 5.01***4.68***3,87**

Right2361014 3.49**5.54**3.60**3.34**3.41**1235679IO14l72.79**3,80***6,OO***2,99*4.27***3,69***2.96**4.18***3.75***2.64***

EP<,05,

**P<.Ol, **-EP<,OOI: _Newman-Keuls

testfoT the diifferencc

in

mean a-cores

between

tbe story and the shape conditions.

nel

1

in

the

left

_prefrontalregion was robust across

tasks inwhich participantshad totryto understand

content. The locusof Channel 1in

the

left_prefrontal

region approximate]y corresponds to

F7

in

the

inter-national 1O!20 system. F7 overlaps theleftinferior

frontal_gyrus and Brodmann's area

(BA)

47

(Okamoto

et al,,

2004>.

BA

47

is

adjacent to,

but

does

not overlap, Breca's area

<BA

44 and 45}. The results of

Experiment

2B

suggest that activation of

Breca's

area, shown as ChanneLs 4 and 5 in the lcft

hemi-sphere

(identified

by

]inearinterpolation,based on

Okamoto

et al.

(2004)),

was

decreased

by

artjculatory suppression. Insuua the increase in

Oxy-Hb

concen-tratlonsinthe leftinieriorfrontalgyrus

(correspond-jng toF7 and BA 47)appears tobea reliab]e measure

to

evaluate whether _participants actively tried

to

understand the content, regardless of the type of

task,rnemory

load,

or the use of

inner

speech.

The

present resu]ts

indicating

thatthetask

involv-ing comprehension activated the leftinferiorfrontal

region

{corresponding

to BA 44/45f47), whereas the

task with requiring articulatory suppression

acti-vated only the ventral region

(BA

47),are consistent with

brain

imaging

studies showing that

pbonologi-nA=,> × o"o98rp5toaoct AD= 1ptxo6y8v5eoavnf 6

5

4

3

2

1

o-1-2

6

5

4

3

2

1 o-1-2

-6

-5

-4

-3

-2

-1

_O

₁

Non-report

(z-score

of

Oxy-Hb)2

-6

-5

-4

-3

-2

-1

O

1

2

Scrembled

(z-score

of Oxy-Hb)

Figure 9. Sca'tter _plot of mean changes in

Oxy-Hb

during the comic presentation in

Experiment 1. The upper _panel shows the

relationship between the report and

report conditiens. The lower _panel shows

the relationship

between

the report and

scrambled conditions. Each circle indicatesa

participant; the

dashed

line

indicates equal

NII-Electronic Library Service

N.

ENDo,

M

NAGAi,

and

T.Kun{ADA:

Objective

Estimation

oi

Understanding

by

NIRS 15

40

330Eito.-20oy8wo 10ul40",en

o

-10

-20

-10

O

10

20

Shape

(z-score

of

Oxy-Hb)

Figure10.

Scatter

plot of mean changes

in

Oxy-Hb during card sorting inExperimcnts

2A and 2B. Each circle or square

indicates

a

participant; the

dashed

line

indicates

equal

brain activation under

both

eonditions.

the

phono]ogical system, A similar dissociationwas

found

in

sentence precessing,

Dapretto

and

Book-heimer

Cl999>

reported that BA 44 was activated

in

processing of syntactic information, whereas BA 47

was selectively

involved

in

processing of the

seman-ticaspects of a sentence, Tlteseresults converged

in

that the

dorsal

part of inferiorfrontal _gyrus

(in

BA

47)

is

involved

in

the _processing of semantic

infor-mation, Thus, our result revealed that BA 47 was

also activated

by

semantic _processing of non-verbal

materials.

In the present study, we analyzed Hb

concentra-tionchanges

during

each o'fthesti,mulation and the

task periods with regard to hemodynarnic response

properties. Although we found more channels

ac-tivated

in

the condition requiring comprehension

during the task period than during the stimulation

period, the specific brain activity in

Channel

1inthe

left

prefrontal region was consistently even

in

each

analysis, To confirm that NIRS activity in the left

inferior

frontal

gyrus

is

a u$efui measure

'to

evaluate

whether participants triedtounderstand, we

exam-ined mean Oxy-Hb concentrations inChannel

1

in

the

le'ft

hemisphere

for

individual

participants

be-tween the task conditions

(Figures

9

and

10).

Most

participants showed higher activation inthe

condt-tion

that

required actjve understanding than inthe

control condition. In

fact,

if

the probe attachment to the sca!p is_good enough toobtain NIRS signals with

a

high

signal-to-noise ratio,we may

be

able

to

obtain

the same

quality

of

data

as seen

in

Figures

9

and

1O

using only one

trial.

If

significant

brain

activation is

observed only

in

a single trial,

it

might

be

possibleto

develop

our procecluretoan online evaluation based

on one trial's

data

and without requiring off-line

analysis of _{participants'}active understanding inthe

future.

A

greater variety of tasks,such as complex textcernprehension and _{proofreading,} could extend

the

areas of application for

this

Qnline measure of

understanding.

For

example, the mea$ure may

be

useful todevelop instruction manuals thatare more readily understood.

References

Dapretto,

M.,

&

Bookheimer,

S.

Y,

(1999).

Form

and

content:

dissociating

syntax and semantics

in

tence comprehension. IVkeuron,24,427-432.

Endo, N.,Saiki,

J.

Nakao, Y. & Saito,H.

_{2003).

ceptual

judgrnents

of novel contour shapes and

hierarchical

descriptions

of _{geometrical properties.}

JtzPanese

LJburnal

of

Psychology,

74,

346-353.

Hoshi,

Y.,

Kobayashi,

N.

&

Tamura. M.

(2001).

pretation of near-infrared spectroscepy signals: a

$tudy with a newly

developed

_perfused rat

brain

model.

fburnal

of

APplied

UPiJJsiotogy,

90,

1662.

Matsuda,

G.

&

Hiraki,

K.

{2006).

Sustained

decrease

in

oxygenated hemoglobin during video _games in

the dorsalprefrontalcortex: a NIRS study of

dren,

Nteuroimage,

29,

706-711.

Okamoto, M. Dan, H.

Sakamoto,

K,,Takeo,

K.

mizu,

K.

Kohno,

S.,

Oda,

I,,

Isobe,

S.

Suzuki,

T,,

Kohyama,

K.

&

Dan,

I.

(2004).

Three-dimensional

probabilistic anatomical cranio-cerebral

tion via the international 10-20 system oriented

for transcranial function brain mapping.

image, 21,99-111.

Oldfield,

R.

C,

{1971),

The

assessment and analysis of

handedness: the Edinburgh inventory.

chologia, 9,97-113.

Otsuka,

Y. Nakato, E. Kanazawa, S.,Yamaguchi, M

K.

Watanabe,

S.

&

Kakigi,R.

{2e07),

Neural

tiontoupright and inverted facesininfants

ured by near infrared spectroscopy. IVeuroJmage,

34,

399-406,

16

The

japanese

Journa]

of

Psychonomic

Science

VoL

28,

No.

1

Peldrack,

R.

A.

Wagner,

A,

D.

PrulL

M.

W.

Desmond,

J.

E.

Glover,

G.

H,,

&

Gabrieli,

J.

D.E.

(1999).

tional specialization for$emantic and _phonological

processing

in

the

left

inferior

prefrontal cortex.

Neuroinzage,

10,

15-35.

Shimada, S. & Hiraki,K,

_(2006).

Infants'brain

sponse$ to livea"d Lelevised action. AIeuroimage,

32,

930-939.

Strangman,

G.

Culver,

J.

P. Thompson,

J.

H,,

&

Boas,

D.A.

(2002).

A quantitatjve comparison of

neous

BOI.D

fMRI

and

NIRS

recordings

during

functional brain activation. IVlauroimage,17,

731,Villringer,

A.,

&

Dirnagl,

U.

(1995).

Coupling

ot

brain

activity and cerebral

blood

fiow:

basis

of

tional neuroimaging. Cerebrovascular and Brain

Metabolism Reviews, 7,240-276.