Developmental Cognitive Neuroscience 1 (2011) 153–162

Contents lists available at ScienceDirect

Developmental Cognitive Neuroscience

journal homepage: http://www.elsevier.com/locate/dcn

Longitudinal development of prefrontal function during early

childhood
Yusuke Moriguchi a,b,∗ , Kazuo Hiraki c,∗
a
Department of School Education, Joetsu University of Education, Japan
b
Japan Science and Technology Agency, PRESTO/Sakigake, Japan
c
Interfaculty Initiative in Information Studies, University of Tokyo, Japan

a r t i c l e i n f o a b s t r a c t

Article history: This is a longitudinal study on development of prefrontal function in young children.
Received 28 July 2010 Prefrontal areas have been observed to develop dramatically during early childhood. To
Received in revised form
elucidate this development, we gave children cognitive shifting tasks related to prefrontal
24 December 2010
function at 3 years of age (Time 1) and 4 years of age (Time 2). We then monitored devel-
Accepted 24 December 2010
opmental changes in behavioral performance and examined prefrontal activation using
near infrared spectroscopy. We found that children showed better behavioral performance
Keywords:
and significantly stronger inferior prefrontal activation at Time 2 than they did at Time 1.
NIRS
Prefrontal cortex Moreover, we demonstrated individual differences in prefrontal activation for the same
Longitudinal study behavioral tasks. Children who performed better in tasks at Time 1 showed significant acti-
Cognitive shifting vation of the right inferior prefrontal regions at Time 1 and significant activation of the
Children bilateral inferior prefrontal regions at Time 2. Children who showed poorer performance
at Time 1 exhibited no significant inferior prefrontal activation at Time 1 but significant
left inferior prefrontal activation at Time 2. These results indicate the importance of the
longitudinal method to address the link between cognitive and neural development.
© 2011 Elsevier Ltd. All rights reserved.

1. Introduction first few years after birth. Numerous studies have shown
that cognitive development is indeed correlated with brain
Over the last few decades, cognitive neuroscience development during infancy and early childhood (Liu et al.,
research has revealed correlations between performance 2009; Moriguchi and Hiraki, 2009; Rueda et al., 2005; Taga
in cognitive tasks and brain activation in specific regions, et al., 2003; Tsujimoto et al., 2004).
such as the prefrontal regions. Various methodological Most developmental cognitive neuroscience studies
approaches have been used, including neuropsycholog- have relied on a cross-sectional design in which an inves-
ical, electrophysiological, and neuroimaging techniques tigator observes several age groups simultaneously. This
(Dias et al., 1996; Konishi et al., 1998; Milner, 1963). approach has several advantages (e.g., it is easier to conduct
Recently, there has been growing interest in the relation- the research), but it cannot be used to determine dynamic
ship between cognitive and brain development during the developmental changes (Kraemer et al., 2000; Magnusson
and Casaer, 1993). Cross-sectional research can clarify the
differences in brain activation and cognitive performance
between two different age groups, but it cannot address
∗ Corresponding authors at: Hiraki Laboratory, Department of General
how the differences occur and whether there are single or
Systems Studies, The University of Tokyo, 3-8-1 Komaba, Meguro, Tokyo
multiple paths of brain and cognitive development. There-
153-8902, Japan. Tel.: +81 3 5454 6994; fax: +81 3 5454 6994.
E-mail addresses: [email protected] (Y. Moriguchi), fore, longitudinal approaches are needed to address these
[email protected] (K. Hiraki). issues. Nevertheless, few longitudinal studies have been

1878-9293/$ – see front matter © 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.dcn.2010.12.004
154 Y. Moriguchi, K. Hiraki / Developmental Cognitive Neuroscience 1 (2011) 153–162

conducted in cognitive neuroscience research, especially cognitive shifting and theory of mind (Garon et al., 2008;
in infants and young children, despite the understanding Liu et al., 2009; Wellman et al., 2001; Zelazo and Müller,
that the early phases of maturation during fetal develop- 2002).
ment and childhood are the most dramatic and important We used the dimensional change card sort (DCCS),
(Toga et al., 2006). which is widely used to index cognitive shifting in young
The present study aimed to investigate longitudinal children (Garon et al., 2008; Zelazo et al., 1996). The task
brain and cognitive development in young children, focus- was chosen because it has been repeatedly shown that chil-
ing on the prefrontal cortex. The prefrontal cortex is a dren improve their performance on the DCCS task between
brain region that subserves complex cognitive functions, 3 and 4 years of age (Kirkham et al., 2003; Moriguchi and
such as cognitive shifting (Crone et al., 2006; Dias et al., Itakura, 2008; Zelazo et al., 1996). In the DCCS task, children
1996; Milner, 1963). Brain imaging studies in adults have are asked to sort cards that have two dimensions, such as
shown that participants were likely to recruit inferior and color and shape. In the preswitch phase, children are asked
dorsolateral prefrontal regions during cognitive shifting to sort cards (e.g., red cups, blue stars) into trays with target
tasks, such as the Wisconsin Card Sorting Test (WCST) cards (e.g., a blue cup, a red star) according to one rule (e.g.,
(Konishi et al., 1998; Monchi et al., 2001; Sumitani et al., color). In the postswitch phase, children are asked to sort
2006). However, recent studies have suggested that there the cards according to a second rule (e.g., shape). Typically,
are some individual differences in the laterality of acti- 3-year-old children perseverate to the first rule, whereas
vation during the WCST. In a near-infrared spectroscopy 4- and 5-year-old children do not (Kirkham et al., 2003;
(NIRS) study, some participants recruited the left pre- Moriguchi et al., 2007; Zelazo et al., 1996).
frontal areas during the WCST, whereas other participants Our previous cross-sectional neuroimaging study using
showed right prefrontal activations (Sumitani et al., 2006). NIRS with the DCCS task showed that 5-year-old children
Moreover, neuropsychological research has revealed that activated bilateral inferior prefrontal regions during the
patients with right lateral prefrontal damage as well as preswitch and the potswtich phases compared to a con-
those with left lateral prefrontal damage showed impaired trol phase, where children simply sorted blank cards into
performance in the WCST task, but the latter showed less a tray (Moriguchi and Hiraki, 2009, but see also a func-
impairment (Stuss et al., 2000). Given this evidence, we tional magnetic resonance imaging [fMRI] study, Morton
believe that individual differences may exist in the later- et al., 2009). Moreover, the NIRS study showed that 3-
ality of prefrontal activations in young children and that year-old children who committed perseverative errors
these differences may be related to individual differences (hereafter, the perseverate group) failed to activate the
in behavioral performance on cognitive tasks. inferior prefrontal regions, whereas those who performed
Recently, extensive research has shown that the the tasks correctly (hereafter, the pass group) activated the
prefrontal cortex develops dramatically during early child- right inferior prefrontal regions during the preswitch and
hood. Behavioral studies have repeatedly shown that postswitch phases. These results suggested that sustained
during preschool years, children’s performance improves right inferior prefrontal activations would be important for
in neuropsychological batteries related to prefrontal func- successful cognitive shifting during the DCCS tasks.
tion (Gerstadt et al., 1994; Luciana and Nelson, 1998; Zelazo The present longitudinal study aimed to extend the
et al., 1996). Moreover, there is some anatomical evi- findings from our cross-sectional study in two ways.
dence that the prefrontal cortex develops during preschool First, we examined the longitudinal development of pre-
years (Diamond, 2002; Giedd et al., 1999; Gogtay et al., frontal activations during the preschool age. Although
2004; Huttenlocher, 1990). Electrophysiological studies our previous cross-sectional study showed differences in
have shown that in children, the electroencephalogram behavioral performance and prefrontal activations dur-
pattern changes between the ages of 1 and 4 years in a ing the DCCS tasks between 3- and 5-year-old children,
working memory task (Bell and Wolfe, 2007) and between it remained unclear whether the behavioral develop-
the ages of 4 and 6 years in a flanker task (Rueda et al., ment was correlated with brain activation. Specifically, the
2005). In addition, in a recent study, neuroimaging data present longitudinal study examined whether brain activa-
have demonstrated functional development of the inferior tion was changed when children’s behavioral performance
prefrontal regions during preschool years in a cognitive improved in the DCCS tasks. Second, individual differences
shifting task (Moriguchi and Hiraki, 2009). in brain activation were also examined. Our previous study
Although behavioral, anatomical, and electrophysiolog- showed that children in the pass group exhibited right infe-
ical research studies have included longitudinal evidence, rior prefrontal activations during the DCCS tasks, whereas
no longitudinal neuroimaging data exist for early child- children in the perseverate group failed to activate inferior
hood (cf., school-aged children, Durston et al., 2006), and no prefrontal regions. Thus, we examined how inferior pre-
research has examined whether there are single or multi- frontal activation changed in children from both groups
ple developmental paths for prefrontal function. Therefore, while performing the same cognitive tasks.
in the present study, we conducted a longitudinal study of In the present study, we gave children the DCCS tasks
the development of the prefrontal function. Specifically, the and examined the developmental changes in prefrontal
present study focused on the longitudinal development of activations using an NIRS technique. Children were tested
prefrontal function between 3 and 4 years of age. We chose twice: at 3 years of age (Time 1) and at 4 years of age (Time
this age period because there is ample evidence that chil- 2). Given the previous NIRS results (Moriguchi and Hiraki,
dren of those ages develop higher order cognitive functions 2009), the region of interest was located at around F7/8
that may be closely related to prefrontal function, such as on the International 10/20 system (Fig. 1A), which corre-
 Y. Moriguchi, K. Hiraki / Developmental Cognitive Neuroscience 1 (2011) 153–162 155

Fig. 1. Experimental settings. (A) The NIRS probe was attached to the inferior prefrontal area. Each channel consisted of one emitter optode and one
detector optode. The region of interest was located near F7/8, which corresponds to channels 15, 17, and 18 and channels 6, 7, and 9, respectively. (B) A
sequence of the experiment.

sponds to Brodmann areas 45/47 (Okamoto et al., 2004). The data at Time 1 were partly reported in Moriguchi and
Brain activation during the DCCS tasks was compared to Hiraki (2009).
activation during the control phases. In the control phases,
children were provided blank cards and instructed to place
the cards into an extra tray. 2.2. Behavioral tasks

Children participated in the experiment sitting on the

2. Materials and methods floor. The behavioral tasks at Time 1 and Time 2 were
almost the same except that different stimuli were used.
2.1. Participants Laminated cards (3.5 cm × 7.0 cm) that had two dimensions
(shape and color) were used as stimuli. The task included
Fifteen right-handed 3-year-old children participated in target cards and test cards; target cards matched test cards
this longitudinal study, but two girls were excluded from in one dimension but did not match in the other dimension
the analyses because they did not participate in the sec- (e.g., a red pig, a green tree, red trees and green pigs. For
ond testing. The mean age at first testing was 41.1 ± 4.0 details, see Appendix A). The experiments included eight
months (mean ± standard deviation [SD]; age range, 37–47 pairs of target and test cards, each of which was different
months; six girls), and the average age at the second test- in shape and color. Four pairs of cards were used at Time 1,
ing was 50.1 ± 4.1 months (age range, 46–55 months). The and the rest of the cards were used at Time 2. There were
study was conducted in accordance with the principles eight pairs of target trays, and each of the trays contained
of the Declaration of Helsinki. Parents provided written target cards. At each session, a different tray with a different
informed consent and were informed verbally of the pur- set of cards was used. For the control task, blank cards and
pose of the study and the safety of the NIRS experiment. The an extra tray were used. The extra tray was placed between
experiments were approved by the local ethics committee. two target trays (Fig. 1B). Basically, the task was children-
156 Y. Moriguchi, K. Hiraki / Developmental Cognitive Neuroscience 1 (2011) 153–162

paced, but the experimenter controlled the children’s pace in our Supporting Information online (Figs. S1 and S2).
when providing the cards. Raw HbO2 data from individual channels without motion
Children were given four consecutive test sessions at artifacts were low-pass filtered through a Savitzky-Golay
Time 1 and four sessions at Time 2. One session consisted filter (Savitzky and Golay, 1964). Following previous stud-
of a control phase (control 1), a preswitch phase, a second ies (Matsuda and Hiraki, 2006; Moriguchi and Hiraki, 2009),
control phase (control 2), and a postswitch phase. Dur- the raw data were converted into Z scores. The Z score was
ing the control phases, children were provided with blank calculated using the mean value and the SD of the changes
cards and asked to place these cards into an extra tray (e.g., in HbO2 concentration during the control phases. Conse-
“All the cards here”). During the preswitch and postswitch quently, the mean value and SD during the control phases
phases, children were given instructions regarding the were respectively changed to Z scores of 0 and 1 in every
rules (e.g., “This is a shape game. All the pigs go here, and channel.
all the trees go there.”), and were asked to sort the cards. Since children were given instructions regarding the
The rule order (color vs. shape first) was constant across rules during the first 5.5 s of the task phases, the last 19.5 s
the four sessions for each child, but the order was coun- during the preswitch and postswitch phases (5.5–25 s)
terbalanced across children. Moreover, the rule order was were analyzed. Average changes in HbO2 during the task
constant across Time 1 and Time 2. In the experiment, phases were calculated for all channels and each subject.
the preswitch and postswitch phases were 25 s each; this The significance of the differences between the changes in
included the time required to provide instructions concern- HbO2 for the baseline (the last 5 s of the control phases)
ing the rules (the first 5.5 s in each phase). Each control and the task (preswitch or postswitch) was determined by
phase was 25 s. a two-tailed Student’s t test for each channel. We com-
The percentage of correct responses and the reaction pared the preswitch phase to control 1 and compared the
times were analyzed. The reaction time was defined as the postswitch phase to control 2. As multiple comparisons
average time taken by the participants to sort one test card. were conducted, we applied a .001 alpha level of signifi-
The reaction time was analyzed with the video tapes. We cance (two groups, two times, two tasks, and six channels).
recorded the reaction time for each trial of each participant. The differences between the pass group and the perse-
verate group as well as between Time 1 and Time 2 (i.e.,
2.3. NIRS recordings and analysis comparisons of the activations during the preswitch phases
at Time 1 to those at Time 2) for each channel were com-
NIRS is a technique for monitoring cerebral hemo- pared using a paired t test, and the significance level was
dynamics by measuring changes in the attenuation of set at p < .001.
near-infrared light passing through tissue (Firbank et al., Second analyses were conducted by comparing the last
1998; Hoshi et al., 2001). The NIRS system used and the 5 s of the task phases to the last 5 s of the control phases
analyses were exactly the same for Time 1 and Time 2. A at both Time 1 and Time 2. We conducted the analyses to
multichannel NIRS unit operating at wavelengths of 780, control the differences of the durations between the con-
805, and 830 nm (OMM-1080S; Shimadzu, Kyoto, Japan) trol phases (5 s) and task phases (19.5 s) in the first analyses.
was used to measure temporal changes in the concentra- The second analyses revealed the basically consistent pat-
tions of oxyhemoglobin (HbO2 ), deoxyhemoglobin (HbR), terns of the brain activations to those of the first analyses,
and total hemoglobin (HbT). One NIRS probe included eight although the effects were somewhat stronger in the second
optodes that comprised 10 channels (Fig. 1A). Each channel analyses than in the first analyses. Thus, in the next section,
consisted of one emitter optode and one detector optode we reported the results of the first analyses.
located 2 cm apart. The sampling rate at each channel was
approximately 10 Hz. 3. Results
The region of interest was located at around F7/8 on the
International 10/20 system (Fig. 1A), which corresponds to The behavioral results as well as the NIRS results at
Brodmann areas 45/47 (Okamoto et al., 2004). Given the Time 2 were directly compared with those at Time 1 to
low spatial resolution of NIRS, channels 6, 7, and 9 (right assess the longitudinal developmental change in cognitive
inferior prefrontal area) and channels 15, 17, and 18 (left shifting and prefrontal activation in young children. Chil-
inferior prefrontal area) were roughly regarded as F8 and dren were classified into the pass group or the perseverate
F7, respectively (Moriguchi and Hiraki, 2009). The test ses- group, according to whether they committed perseverative
sions in which motion artifacts were revealed by the video errors during at least one of the four assigned sessions at
recordings and NIRS data were discarded. Variations for Time 1. In other words, even if a child committed persever-
each sample datum of HbO2 were calculated by subtract- ative errors in the first session, but not in other sessions,
ing a previous datum from a current datum. Channels in the child was classified into the perseverate group. The
which a variation of more than 3 SDs was detected were mean number of the sessions the children committed the
excluded from further analysis. We rejected the trial where perseverative errors was 2.0 (SD = 1.26, range 1–4).
an artifact was detected. Ultimately, approximately 6% of
the data from the children were excluded from the anal- 3.1. Behavioral results
yses. Of the three NIRS parameters measured, a change
in the HbO2 concentration was considered to be the best The mean percentage of correct responses by children
indicator of brain activity (Hock et al., 1997; Toronov et al., at Time 1 was 94.9% in the preswitch phases and 72.1%
2001). For the validity, the other parameters were reported during the postswitch phases. Six of the 13 children were
 Y. Moriguchi, K. Hiraki / Developmental Cognitive Neuroscience 1 (2011) 153–162 157

Fig. 2. Behavioral results. (A) Percentage of correct responses and (B) reaction time during the preswitch and postswitch phases at Time 1 and Time 2.
Error bars indicate standard error (SE). * p < .05.

classified into the perseverate group, and the remaining dren activated the right inferior prefrontal regions (channel
7 were placed in the pass group. At Time 2, on average, 9) during the preswitch phase but failed to activate the
children performed correctly 96.4% during the preswitch bilateral inferior prefrontal areas during the postswitch
phases and 98.9% during the postswitch phases (Fig. 2A). phase as compared with control phases (Student’s t test,
None of the children committed perseverative errors at p < .001). At Time 2, in the right (channels 6 and 7) and
Time 2. The mean percentage of correct responses for the left (channels 15, 17, and 18) inferior prefrontal areas, the
pass group and the perseverate group was 99.0% and 90.3% mean change in HbO2 was significantly higher during the
during the preswitch phases and 96.6% and 43.2% dur- preswitch phase than during the control phases. During the
ing the postswitch phases at Time 1, respectively. Also, postswitch phase, children exhibited significant changes
the mean reaction time for the pass group and the per- in HbO2 in the right (channel 9) and left (channels 15,
severate group at Time 2 was 97.9% and 94.7% during 17, and 18) inferior prefrontal areas (p < .001). We then
the preswitch phases and 97.9% and 100.0% during the directly compared brain activation at Time 1 with that at
postswitch phases, respectively. The percentage of correct Time 2 and found that children showed significant activa-
responses was analyzed using a 2 (Time 1 vs. Time 2) × 2 tion in the right and left inferior prefrontal areas during the
(preswitch vs. postswitch) ANOVA, and we observed a sig- preswitch (channels 6 and 7 and channels 15, 17, and 18)
nificant interaction of the percentage of correct responses and postswitch phases (channels 6, 7, and 9 and channels
[F (1, 95) = 11.258, p < .001, 2 = .11]. 15, 17, and 18) at Time 2 as compared with Time 1 (paired
Furthermore, we compared reaction times at Time 1 t test, p < .001). Further analyses conducted to explore the
with reaction times at Time 2 and found that children difference between preswitch and postswitch phases are
required more time to sort the cards during the postswitch reported in our Supporting Information online.
phase (M = 3.5 s) than during the preswitch phase (M = 2.7 s)
at Time 1, whereas the reaction time at Time 2 was almost
the same during both preswitch (M = 3.0 s) and postswitch
phases (M = 3.0 s) (Fig. 2B). The reaction times for the
pass group and the perseverate group were 2.3 s and 3.3 s
during the preswitch phases and 2.8 s and 4.5 s during
the postswitch phases at Time 1, respectively. Also, the
mean reaction time for the pass group and the persever-
ate group at Time 2 was 2.7 s and3.3 s during the preswitch
phases and 2.7 s and 3.2 s during the postswitch phases,
respectively. We conducted a 2 (condition; pass vs. perse-
verate) × 2 (time; Time 1 vs. Time 2) × 2 (phase; preswitch
vs. postswitch) ANOVA and found a main effect of condi-
tion [F (1, 86) = 17.128, p < .001, 2 = .13] and phase [F (1,
86) = 5.675, p < .02, 2 = .06] as well as a significant inter-
action between time and phase [F (1, 86) = 6.292, p < .02,
2 = .07]. Both accuracy and efficiency results therefore
showed that children improved their behavioral perfor-
mance of the cognitive shifting task between Time 1 and
Time 2.

3.2. NIRS results

Fig. 3. Data from children in the whole group at Time 1 (A, B) and Time
3.2.1. Whole-group analysis 2 (C, D) during the DCCS task. Averaged overall data during the task
phases compared with the control phases are shown. The numbers (1–20)
First, we examined whether there was a developmental
indicate the channels of the NIRS probe. (A) Preswitch phase and (B)
change in brain activation for the whole group (Fig. 3). At postswitch phase at Time 1. (C) Preswitch phase and (D) postswitch phase
Time 1, the results of the NIRS recordings revealed that chil- at Time 2.
158 Y. Moriguchi, K. Hiraki / Developmental Cognitive Neuroscience 1 (2011) 153–162

3.2.2. Separate analyses ited both right and left inferior prefrontal activations at
Next, we analyzed the children separately according to Time 2.
the perseverative errors observed at Time 1. First, we exam- Second, we examined whether children in the per-
ined whether children in the pass group (n = 7) showed severate group showed different activations at Time 2
similar or different brain activations for Time 1 and Time compared with Time 1. At Time 1, children in the per-
2. In terms of similarity in brain activation, children in the severate group exhibited no significant HbO2 increases
pass group showed significant right inferior prefrontal acti- in either the right or left inferior prefrontal areas dur-
vation (channel 6) during the preswitch and postswitch ing the task phases compared with the control phases.
phases at Time 1 and Time 2 as compared with the con- Rather, they showed deactivations across the preswitch
trol phases (p < .001; Figs. 4 and 5). In terms of differences (channels 6, 15, and 18) and postswitch phases (channels
in brain activation, children in the pass group activated the 6, 9, and 15) (Figs. 4 and 6A,B). However, at Time 2, they
left inferior prefrontal areas during the preswitch (channel showed significant increases in HbO2 in the left inferior
15) and postswitch phases (channels 15, 17, and 18) at Time prefrontal regions during the preswitch (channels 17 and
2 but not at Time 1 as compared with the control phases 18) and postswitch phases (channels 15 and 18) (p < .001)
(Figs. 4 and 5). We then directly compared the brain activa- compared with the control phases. They did not exhibit
tions at Time 1 with those at Time 2 and found that children significant HbO2 increases in the right inferior prefrontal
in the pass group exhibited greater changes in HbO2 in areas during either phase, although task phases seemed to
the inferior prefrontal areas during the preswitch phase affect the brain activations (Figs. 4 and 6C,D). Next, we com-
(channels 6, 7, and 15) and postswitch phase (channels 6 pared the brain activations at Time 1 with those at Time 2
and 7 and channels 15 and 18) at Time 2 than they did at directly and found that children in the perseverate group
Time 1 (p < .001) (Fig. S3). These results revealed that during exhibited significant HbO2 increases in the bilateral infe-
the DCCS tasks, children in the pass group exhibited right rior prefrontal areas during the preswitch phase (channels
inferior prefrontal activations at Time 1, while they exhib- 6 and 7 and channels 15, 17, and 18) and the postswitch

Fig. 4. Temporal changes in the HbO2 concentration in the right (channel 6) and left (channel 15) inferior prefrontal areas during the experiment. (A) and
(B) show the preswitch phases and the postswitch phases, respectively. Data of the group mean in children in the pass group at Time 1 (yellow line) and
Time 2 (red line), and children in the perseverate group at Time 1 (green line) and Time 2 (blue line) are shown. (For interpretation of the references to
color in this figure legend, the reader is referred to the web version of the article.)
 Y. Moriguchi, K. Hiraki / Developmental Cognitive Neuroscience 1 (2011) 153–162 159

Finally, we directly compared the brain activation in the

pass group with those in the perseverate group. At Time 1,
children in the pass group exhibited significantly stronger
inferior prefrontal activations than did those in the perse-
verate group during the preswitch phase (channels 6, 7 and
9 and channels 15, 17, and 18) and the postswitch phase
(channels 6, 7, and 9 and channels 15 and 17) (p < .001).
Moreover, at Time 2, children in the pass group showed
significantly greater activations compared with the perse-
verate group during the preswitch phase (channels 6, 7 and
9 and channels 15, 17, and 18) and postswitch phase (chan-
nel 17), while also showing significant deactivations during
the postswitch phases (channel 9) (p < .001).

4. Discussion

Our longitudinal study clarified changes in behav-

ioral performance and brain development during cognitive
shifting occurring in young children. On the behavioral
level, children showed a significant improvement in cog-
nitive shifting as measured by the DCCS tasks. Six children
Fig. 5. Data from children in the pass group at Time 1 (A, B) and Time committed perseverative errors at Time 1, but none of the
2 (C, D) during the DCCS task. Averaged overall data during the task children committed errors at Time 2. On the neural level,
phases compared with the control phases are shown. The numbers (1–20) children showed a significant improvement in inferior pre-
indicate the channels of the NIRS probe. (A) Preswitch phase and (B)
frontal activations between Time 1 and Time 2. At Time 1, in
postswitch phase at Time 1. (C) Preswitch phase and (D) postswitch phase
at Time 2. general, children showed weaker inferior prefrontal activa-
tions across the DCCS tasks; at Time 2, children significantly
phase (channels 6, 7, and 9 and channels 15 and 18) at Time activated bilateral inferior prefrontal regions during the
2 as compared with Time 1 (p < .001) (Fig. S4). These results preswitch phase and left inferior prefrontal regions during
revealed that children in the perseverate group exhibited the postswitch phase.
no significant inferior prefrontal activations at Time 1, but These results support and extend our previous cross-
showed significant left inferior prefrontal activations at sectional findings (Moriguchi and Hiraki, 2009). The
Time 2 during the DCCS tasks. previous study documented significant differences in
behavioral performance and inferior prefrontal activations
during the DCCS tasks between 3- and 5-year-old children.
However, it was still unclear whether the behavioral dif-
ferences were related to the inferior prefrontal activations.
The present study suggests that the behavioral devel-
opment observed during the DCCS tasks was correlated
with the inferior prefrontal activations. The results implied
that the development of cognitive shifting might be sup-
ported by the development of the inferior prefrontal cortex,
although we could not discuss the causal relation between
behavioral and neural changes from the results. More
generally, this study is the first to provide longitudinal
data indicating that children develop prefrontal activations
between the ages of 3 and 4 years. Although some lon-
gitudinal anatomical and electrophysiological researches
have demonstrated the development of prefrontal func-
tion during the preschool years (e.g., Rueda et al., 2005),
little neuroimaging data are available regarding prefrontal
development during this time period. Coupled with the
behavioral, anatomical, and electrophysiological evidence,
the neuroimaging data would contribute to our under-
standing of the development of the prefrontal cortex during
early childhood.
Fig. 6. Data from children in the perseverate group at Time 1 (A, B) and In the present study, children generally showed
Time 2 (C, D) during the DCCS task. Averaged overall data during the stronger left inferior prefrontal activations at Time 2 than at
task phases compared with the control phases are shown. The numbers
Time 1. There are several possible interpretations regard-
(1–20) indicate the channels of the NIRS probe. (A) Preswitch phase and
(B) postswitch phase at Time 1. (C) Preswitch phase and (D) postswitch ing these laterality effects. One possibility is that children
phase at Time 2. developed their hand-preference during the period of the
160 Y. Moriguchi, K. Hiraki / Developmental Cognitive Neuroscience 1 (2011) 153–162

study. That is, children at Time 2 may have tended to in the left inferior prefrontal areas or the “improvement”
use the right hand to sort the cards compared with Time of the deactivations in the right inferior prefrontal regions.
1. However, we refute this possibility because even at In both interpretations, the results for children in the
Time 1, children showed a strong hand preference, and perseverate group at Time 2 were different from the results
all of the children were right-handed and sorted the cards for children in the pass group at Time 1 and Time 2. The
with their right hands. The second possibility is that chil- results for Time 1 and Time 2 suggested that unilateral
dren developed speech perception during the study period, (either right or left) inferior prefrontal activations may be
especially processing of the sentences that communicated important for successful performance in cognitive shifting
the experiment’s instructions. It has been shown in adult tasks, such as the DCCS task. A previous study suggested
fMRI studies that sentence processing may rely on the left that “sustained right inferior prefrontal activation might
frontal, temporal, and parietal areas (e.g., Friederici et al., be crucial for successful cognitive shifting in young chil-
2003). However, the fact that even at Time 1, children per- dren” (Moriguchi and Hiraki, 2009, p. 6019). Instead, we
formed the preswitch phase of the DCCS task perfectly may suggest that sustained inferior prefrontal activations rather
weaken this possibility. In the DCCS tasks, the instructions than sustained right inferior prefrontal activations may
were fundamentally the same in both the preswitch and be crucial for successful cognitive shifting in young chil-
postswitch phases. This suggests that children at Time 1 dren. Taken together, these separate analyses suggest the
understood the instructions for the experiment but failed existence of multiple paths in the development of the pre-
to activate the left inferior prefrontal regions. The third frontal activations observed during the cognitive shifting
possibility is that children may recruit inner speech to per- tasks. However, due to the small sample size in both anal-
form the tasks correctly. Kirkham et al. (2003) showed that yses, these findings are, at present, inconclusive. Further
labeling a new rule in the DCCS task may lead to better research should be done to assess and confirm the present
performance. Also, children may internalize their private findings with a bigger sample size.
speech during early childhood (Manfra and Winsler, 2006; In general, our findings with regard to unilateral acti-
Vygotsky, 1988). This evidence may suggest that children vations seem to be consistent with those of some previous
at Time 2 were more likely to engage inner speech to per- studies in which participants recruited either right or left
form the tasks than they were at Time 1, which may have inferior prefrontal regions during cognitive shifting tasks.
activated the left inferior prefrontal regions. Adult brain imaging studies have suggested that there
One advantage of the longitudinal design is that we are individual differences in the laterality of the activa-
could assess how behavioral development is related to tion during the WCST; some participants recruited the left
brain development. When the children were classified into prefrontal areas, whereas other participants demonstrated
the pass group and perseverate group, analyses yielded right prefrontal activations (Sumitani et al., 2006). Bunge
interesting results. At Time 1, children in the pass group et al. (2002) reported that adult participants recruited
exhibited right inferior prefrontal activations across the right ventrolateral prefrontal regions during a flanker
preswitch and postswitch phases as compared with the task, whereas school-aged children activated left ventro-
control phases. At Time 2, they exhibited activations in lateral prefrontal regions, suggesting that children may use
both the right and left inferior prefrontal regions. The acti- different strategies compared with adults. Moreover, neu-
vation pattern was similar to that of 5-year-old children ropsychological research has revealed that patients with
in a previous study who exhibited bilateral inferior pre- right or left lateral prefrontal damage experienced diffi-
frontal activations during the DCCS tasks (Moriguchi and culty with the WCST, but those with left lateral prefrontal
Hiraki, 2009). It is likely that a developmental process exists damage (i.e., those with intact right prefrontal regions)
whereby children first engage the right inferior prefrontal showed less impairment (Stuss et al., 2000). This neuropsy-
regions and then recruit the bilateral inferior prefrontal chological evidence may accord with the present results
regions during the DCCS task. indicating that 3-year-old children in the pass group first
Nevertheless, the story is not that simple. Children in recruited the right inferior prefrontal areas. In other words,
the perseverate group showed a different developmental right inferior prefrontal areas may be relatively dominant
pattern. At Time 1, children in this group failed to engage in cognitive shifting tasks, and left inferior prefrontal areas
the bilateral prefrontal regions during the preswitch and may support or compensate for right inferior prefrontal
postswitch phases. Rather, their inferior prefrontal regions activations.
exhibited deactivations during the DCCS task. However, at Finally, the fact that there might be multiple neural
Time 2 the children performed the DCCS tasks almost per- developmental paths in prefrontal function may contribute
fectly (as did children in the pass group at both Time 1 to our understanding of developmental disorders such
and Time 2), showing significant left, but not right, inferior as attention-deficit hyperactivity disorder (ADHD). It has
prefrontal activations during the preswitch and postswitch been hypothesized that patients with ADHD may have
phases as compared with the control phases. In terms of the functional deficits in the prefrontal cortex (Bush et al.,
right inferior prefrontal areas, children in the perseverate 2005). However, it is still unclear when and how children
group showed changes between Time 1 and Time 2. At Time with ADHD exhibit this deficit in prefrontal function. Chil-
1, these children showed deactivations in the right infe- dren with ADHD may exhibit a delay in the maturation of
rior prefrontal areas during the DCCS tasks, whereas these prefrontal areas (Shaw et al., 2007). Nevertheless, given
deactivations were not observed at Time 2. We believe that the results of the present study, it may be possible that
the differences in performance of the DCCS tasks between prefrontal function follows a different developmental path-
Time 1 and Time 2 may be due to the significant activations way in early childhood in children with ADHD compared
 Y. Moriguchi, K. Hiraki / Developmental Cognitive Neuroscience 1 (2011) 153–162 161

with children undergoing typical development. Further Friederici, A.D., Ruschemeyer, S.A., Hahne, A., Fiebach, C.J., 2003. The role of
research using longitudinal methods should be performed left inferior frontal and superior temporal cortex in sentence compre-
hension: localizing syntactic and semantic processes. Cereb. Cortex
to address these issues. 13, 170–177.
Garon, N., Bryson, S.E., Smith, I.M., 2008. Executive function in preschool-
ers: a review using an integrative framework. Psychol. Bull. 134,
Acknowledgements 31–60.
Gerstadt, C.L., Hong, Y.J., Diamond, A., 1994. The relationship between
This research was supported by grants from JSPS to cognition and action: performance of children 3.5–7 years old on a
Stroop-like day-night test. Cognition 53, 129–153.
Kazuo Hiraki and Yusuke Moriguchi. Also, this was sup- Giedd, J.N., Blumenthal, J., Jeffries, N.O., Castellanos, F.X., Lui, H., Zijdenbos,
ported by JST PRESTO program. We thank G. Matsuda, R. A., Paus, T., Evans, A.C., Rapoport, J.L., 1999. Brain development during
Matsunaka, and I. Shinohara for help with data collection. childhood and adolescence: a longitudinal MRI study. Nat. Neurosci.
2, 861–863.
We also thank the parents and children who participated
Gogtay, N., Giedd, J.N., Lusk, L., Hayashi, K.M., Greenstein, D., Vaituzis, A.C.,
in this study. Nugent III, T.F., Harman, D.H., Clasen, L.S., Toga, A.W., Rapoport, J.L.,
Thompson, P.M., 2004. Dynamic mapping of human cortical develop-
ment during childhood through early adulthood. Proc. Natl. Acad. Sci.
Appendix A. Appendix U.S.A. 101, 8174–8179.
Hock, C., Villringer, K., Muller-Spahn, F., Wenzel, R., Heekeren, H., Schuh-
Stimuli used at Time 1 and Time 2 Hofer, S., Hofmann, M., Minoshima, S., Schwaiger, M., Dirnagl, U.,
Villringer, A., 1997. Decrease in parietal cerebral hemoglobin oxy-
Time 1 genation during performance of a verbal fluency task in patients
with Alzheimer’s disease monitored by means of near-infrared
spectroscopy (NIRS): correlation with simultaneous rCBF-PET mea-
Pair 1: a red star, a blue cup, red cups, and blue stars. surements. Brain Res. 755, 293–303.
Pair 2: a green car, a yellow house, green houses, and yel- Hoshi, K., Kobayashi, N., Tamura, M., 2001. Interpretation of near-infrared
low cars. spectroscopy signals: a study with a newly developed perfused rat
brain model. J. Appl. Physiol. 90, 1657–1662.
Pair 3: an orange face, a purple flower, orange flowers and
Huttenlocher, P.R., 1990. Morphometric study of human cerebral cortex
purple faces. development. Neuropsychologia 28, 517–527.
Pair 4: a gray camera, a brown computer, gray computers, Kirkham, N.Z., Cruess, L., Diamond, A., 2003. Helping children apply their
knowledge to their behavior on a dimension-switching task. Dev. Sci.
and brown cameras.
6, 449–476.
Konishi, S., Nakajima, K., Uchida, I., Kameyama, M., Nakahara, K., Seki-
Time 2 hara, K., Miyashita, Y., 1998. Transient activation of inferior prefrontal
cortex during cognitive set shifting. Nat. Neurosci. 1, 80–84.
Kraemer, H.C., Yesavage, J.A., Taylor, J.L., Kupfer, D., 2000. How can we
Pair 1: a red pig, a green tree, red trees and green pigs. learn about developmental processes from cross-sectional studies, or
can we? Am. J. Psychiatry 157, 163–171.
Pair 2: a yellow bag, a blue truck, yellow trucks, blue bags.
Liu, D., Sabbagh, M.A., Gehring, W.J., Wellman, H.M., 2009. Neural cor-
Pair 3: a purple triangle, a brown circle, purple circles, and relates of children’s theory of mind development. Child Dev. 80,
brown triangles. 318–326.
Pair 4: a gray cloud, a white TV, gray TVs, and white clouds. Luciana, M., Nelson, C.A., 1998. The functional emergence of prefrontally-
guided working memory systems in four- to eight-year-old children.
Neuropsychologia 36, 273–293.
Appendix B. Supplementary data Magnusson, D., Casaer, P., 1993. Longitudinal Research on Individual
Development: Present Status and Future Perspectives. Cambridge
University Press, Cambridge.
Supplementary data associated with this arti- Manfra, L., Winsler, A., 2006. Preschool children’s awareness of private
cle can be found, in the online version, at speech. Int. J. Behav. Dev. 30, 537–549.
Matsuda, G., Hiraki, K., 2006. Sustained decrease in oxygenated
doi:10.1016/j.dcn.2010.12.004.
hemoglobin during video games in the dorsal prefrontal cortex: a NIRS
study of children. Neuroimage 29, 706–711.
Milner, B., 1963. Effects of different brain lesions on card sorting. Arch.
References
Neurol. 9, 90–100.
Monchi, O., Petrides, M., Petre, V., Worsley, K., Dagher, A., 2001. Wisconsin
Bell, M.A., Wolfe, C.D., 2007. Changes in brain functioning from infancy to card sorting revisited: distinct neural circuits participating in differ-
early childhood. Dev. Neuropsychol. 31, 21–38. ent stages of the task identified by event-related functional magnetic
Bunge, S.A., Dudukovic, N.M., Thomason, M.E., Vaidya, C.J., Gabrieli, J.D.E., resonance imaging. J. Neurosci. 21, 7733–7741.
2002. Immature frontal lobe contributions to cognitive control in chil- Moriguchi, Y., Hiraki, K., 2009. Neural origin of cognitive shifting in young
dren: evidence from fMRI. Neuron 33, 301–311. children. Proc. Natl. Acad. Sci. U.S.A. 106, 6017–6021.
Bush, G., Vaela, E.M., Seidman, L.J., 2005. Functional neuroimaging Moriguchi, Y., Itakura, S., 2008. Young children’s difficulty with inhibitory
of attention-deficit/hyperactivity disorder: a review and suggested control in a social context. Jpn. Psychol. Res. 50, 87–92.
future directions. Biol. Psychiatr. 57, 1273–1284. Moriguchi, Y., Lee, K., Itakura, S., 2007. Social transmission of disinhibition
Crone, E., Donohue, S., Honomichl, R., Wendelken, C., Bunge, S., 2006. Brain in young children. Dev. Sci. 10, 481–491.
regions mediating flexible rule use during development. J. Neurosci. Morton, J.B., Bosma, R., Ansari, D., 2009. Age-related changes in brain acti-
26, 11239–11247. vation associated with dimensional shifts of attention: an fMRI study.
Diamond, A., 2002. Normal development of prefrontal cortex from birth Neuroimage 46, 249–256.
to young adulthood: cognitive functions, anatomy, and biochemistry. Okamoto, M., Dan, H., Sakamoto, K., Takeo, K., Shimizu, K., Kohno, S., Oda,
In: Stuss, D.T., Knight, R.T. (Eds.), Principles of Frontal Lobe Function. I., Isobe, S., Suzuki, T., Kohyama, K., Dan, I., 2004. Three-dimensional
Oxford Univ. Press, New York, pp. 466–503. probabilistic anatomical cranio-cerebral correlation via the inter-
Dias, R., Robbins, T.W., Roberts, A.C., 1996. Dissociation in prefrontal cortex national 10–20 system oriented for transcranial functional brain
of affective and attentional shifts. Nature 380, 69–72. mapping. Neuroimage 21, 99–111.
Durston, S., Davidson, M.C., Tottenham, N., Galvan, A., Spicer, J., Fossella, Rueda, M.R., Rothbart, M.K., Saccamanno, L., Posner, M.I., 2005. Training,
J.A., Casey, B.J., 2006. A shift from diffuse to focal cortical activity with maturation, and genetic influences on the development of executive
development. Dev. Sci. 9, 1–20. attention. Proc. Natl. Acad. Sci. U.S.A. 102, 14931–14936.
Firbank, M., Okada, E., Delpy, D.T., 1998. A theoretical study of the signal Savitzky, A., Golay, M.J.E., 1964. Smoothing and differentiation of data by
contribution of regions of the adult head to near-infrared spectroscopy simplified least squares procedures. Anal. Chem. 36, 1627–1639.
studies of visual evoked responses. Neuroimage 8, 69–78.
162 Y. Moriguchi, K. Hiraki / Developmental Cognitive Neuroscience 1 (2011) 153–162

Shaw, P., Eckstrand, K., Sharp, W., Blumenthal, J., Lerch, J.P., Greenstein, Toronov, V., Webb, A., Choi, J.H., Wolf, M., Michalos, A., Gratton, E., Hueber,
D., Clasen, L., Evans, A., Giedd, J., Rapoport, J.L., 2007. Attention- D., 2001. Investigation of human brain hemodynamics by simultane-
deficit/hyperactivity disorder is characterized by a delay in cortical ous near-infrared spectroscopy and functional magnetic resonance
maturation. Proc. Natl. Acad. Sci. U.S.A. 104, 19649–19654. imaging. Med. Phys. 28, 521–527.
Stuss, D.T., Levine, B., Alexander, M.P., Hong, J., Palumbo, C., Hamer, L., Mur- Tsujimoto, S., Yamamoto, T., Kawaguchi, H., Koizumi, H., Sawaguchi, T.,
phy, K.J., Izukawa, D., 2000. Wisconsin Card Sorting Test performance 2004. Prefrontal cortical activation associated with working memory
in patients with focal frontal and posterior brain damage: effects of in adults and preschool children: an event-related optical topography
lesion location and test structure on separable cognitive processes. study. Cereb. Cortex 14, 703–712.
Neuropsychologia 38, 388–402. Vygotsky, L.S., 1988. On inner speech. In: Franklin, M.B., Barten, S. (Eds.),
Sumitani, S., Tanaka, T., Tayoshi, S., Ota, K., Kameoka, N., Ueno, S., Ohmori, Child Language: A Reader. Oxford University Press, New York, pp.
T., 2006. Activation of the prefrontal cortex during the Wisconsin Card 181–187.
Sorting Test as measured by multichannel near-infrared spectroscopy. Wellman, H., Cross, D., Watson, J., 2001. Meta-analysis of theory-of-mind
Neuropsychobiology 53, 70–76. development: the truth about false belief. Child Dev. 72, 655–684.
Taga, G., Asakawa, K., Maki, A., Konishi, Y., Koizumi, H., 2003. Brain imaging Zelazo, P.D., Frye, D., Rapus, T., 1996. An age-related dissociation between
in awake infants by near-infrared optical topography. Proc. Natl. Acad. knowing rules and using them. Cogn. Dev. 11, 37–63.
Sci. U.S.A. 100, 10722–10727. Zelazo, P.D., Müller, U., 2002. Executive function in typical and atypical
Toga, A.W., Thompson, P.M., Sowell, E.R., 2006. Mapping brain maturation. development. In: Goswami, U. (Ed.), Blackwell Handbook of Childhood
Trends Neurosci. 29, 148–159. Cognitive Development. Blackwell, Oxford, England, pp. 445–469.