Abstract

It is well known that lateral areas of the prefrontal cortex (LPFC) play a central role in working memory (a critical basis of various cognitive functions), but it remains unknown whether the LPFC of children of preschool age is responsible for working memory. To address this issue, we adopted a recently developed non-invasive imaging technique, optical topography (OT), which can potentially be applied to functional mapping in childhood. We firstly examined changes of activity in the LPFC using OT while adult subjects performed an item-recognition task, which requires working memory, under different memory-load conditions. We observed activation in the bilateral LPFC during performance of this task, the magnitude of which differed depending on memory-load. Then, we applied the same technique on 5- and 6-year-old children and observed the activation associated with working memory in the LPFC. Areas and properties of such activity were similar in adults and preschool children. Thus, for the first time, we demonstrate that the LPFC of preschoolers is active during working memory processes, indicating that in 5- and 6-year-old children, the LPFC has already developed processing of this important cognitive function.

Introduction

Working memory is a brain system involved in temporary storage and manipulation of information and is a key requirement for various cognitive functions such as planning and reasoning and, thus, everyday life (Baddeley, 1986, 1992). Study of its development at the level of the brain should be significant for the understanding of neural development associated with this important cognitive function (Huttenlocher, 2003). However, most studies of the neural basis of working memory have examined adults of human or non-human primates. For example, in non-human primates, both lesion and single-neuron studies using a spatial delayed-response task (a kind of working memory task) have demonstrated that lateral areas of the prefrontal cortex (LPFC) contain the neural substrate for visuospatial short-term (or working) memory (Goldman and Rosvold, 1970; Bauer and Fuster, 1976; Funahashi et al., 1989, 1993; Sawaguchi and Yamane, 1999; Sawaguchi and Iba, 2001; for reviews, see Funahashi and Kubota, 1994; Goldman-Rakic, 1995; Fuster, 1997). In addition, in the adult human, many studies using neuropsychological and brain-imaging approaches have provided significant evidence of LPFC involvement in working memory (Freedman and Oscar-Berman, 1986; Jonides et al., 1993; McCarthy et al., 1994; Shimamura, 1994; Owen et al., 1996, 1999; Sweeney et al., 1996; Belger et al., 1998; Courtney et al., 1998; Leung et al., 2002). Thus, the LPFC plays a central role in working memory.

A limited number of developmental studies have demonstrated that working memory performance improves with age from childhood to young adulthood (Kail and Salthouse, 1994; Luciana and Nelson, 1998; Zald and Iacono, 1998). In particular, Luciana and Nelson (1998) reported, using psychological approaches, that prefrontally guided working memory systems emerge at around the age of 4 years and improve substantially between the ages of 5 and 7 years. Furthermore, morphological studies have demonstrated that brain structure changes substantially around preschool age, in terms of the volume of grey and white matter (Giedd et al., 1999) and the number of synapses in the PFC (Huttenlocher, 1979). Therefore, it is important to study the functional development of working memory in childhood, especially around preschool age; such an approach should extend our understanding of not only cognitive development, but also the neural mechanisms of cognitive functions (Casey et al., 2000). However, although some fMRI studies of working memory development have been performed in children aged 7 years or older (Casey et al., 1995; Thomas et al., 1999; Nelson et al., 2000; Kwon et al., 2002), the neural basis of working memory development in preschool children is poorly understood. Particularly, it is still unknown whether, as in adults, the LPFC of preschoolers is active during working memory tasks. This deficit in understanding mainly results from the fact that many of the brain-imaging procedures are invasive or require extensive constraints and thus cannot easily be used to study normally developing children, in particular those of preschool age.

The aim of the present study is to reveal whether the LPFC of preschool children is active during working memory. To address this issue, we adopted a recently developed non-invasive imaging technique, optical topography (OT; Maki et al., 1995; Yamashita et al., 1996; Koizumi et al., 1999, 2003). OT is a type of near-infrared spectroscopy (NIRS) imaging (Jöbsis, 1977) and has several advantages over other imaging methods such as fMRI or PET, e.g. OT does not require a head constraint and hence can potentially be applied to the developmental mapping of children and infants. Previous studies using OT have reported haemodynamic changes associated with various cortical functions such as motor action (Maki et al., 1995, 1996) and language processing (Watanabe et al., 1998; Sato et al., 1999; Kennan et al., 2002). Further, Noguchi et al. (2002) employed OT to examine event-related changes in activity of the cerebral cortex, including LPFC, during syntactic and semantic decision tasks and demonstrated that OT is useful for studying the higher cognitive functions, although studies using OT on LPFC-mediated cognitive functions have been limited.

In this study, we firstly examined changes in activity in the LPFC of human adults using OT with an event-related paradigm, while subjects performed an item-recognition task, which requires working memory, under high and low memory-load conditions. Then, we used the same technique with preschool children (aged from 5 to 6 years). We report here that the LPFC not only of adults, but also of preschool children, is active during working memory, and the region and properties of such activity are similar in adults and preschool children.

Materials and Methods

Subjects

Seven right-handed healthy adults [aged between 22 and 28 years, 25.4 ± 1.9 (mean ± SD), six males and one female] and 16 healthy right-handed preschool children [aged between 5 and 6 years, 69.7 ± 6.7 months (mean ± SD), six males and 10 females] participated in this study. The children attended Shiratori Kindergarten, Saitama, Japan. All adult subjects provided informed consent for this study. For the children, parents provided written informed consent and were informed verbally of the purpose of the study and the safety of the OT experiment.

Behavioural Task

The behavioural task for adults was an item-recognition task with two memory-load conditions (i.e. HIGH and LOW conditions; see Fig. 1), which was controlled by a personal computer (PCG-505; Sony, Japan). A trial commenced when the central fixation spot (grey cross, 1 × 1°) turned to white (warning signal). After a warning period of a variable interval (1–3 s), two or four white squares (1.5 × 1.5°) were presented at the peripheral locations (two or four of eight locations) in the LOW and HIGH conditions, respectively (sample cue, 2 s), which was followed by a delay period of 8 s. Then, a white square (1.5 × 1.5°) was presented as a test cue at one of eight peripheral locations and, based on working memory of the locations of the sample cue, the subjects were required to report whether the location of the test cue was identical to the location of any of the sample cues. The response was reported by pressing a button: ‘yes’ was indicated by the right index finger and ‘no’ by the left index finger. When the subject responded, the test cue was turned off and the trial ended. A trial was followed by an intertrial interval (ITI) of 25 s, where a grey cross was presented in the centre of the display. Throughout the trial, the subjects were instructed to maintain fixation on the central cross. ‘Yes’ and ‘no’ trials were pseudo-randomized to be the same number (i.e. five trials for ‘yes’ and the other five for ‘no’).

Each adult subject performed two blocks of trials, each consisting of 10 LOW or HIGH trials. The order of conditions was counter-balanced among the seven subjects. Each child performed 10 trials for the LOW condition only, because of their difficulty in performing, or poor performance in, the HIGH condition. Children received sufficient training before the measuring session so that they were able to perform the task adequately (i.e. >80% correct response).

Optical Topography Measurements

While the subjects performed the behavioural task, we measured haemoglobin oxygenation in the LPFC using an OT system (ETG-100; Hitachi Medical Corporation, Tokyo, Japan). Near-infrared laser diodes with two wavelengths (780 and 830 nm) were used as light emitters. The re-emitted lights were detected with avalanche photodiodes located 30 mm from the emitters. In the LPFC of each hemisphere, five emitters and four detectors were placed at alternate square intersection points on a 3 × 3 grid (Fig. 2). This configuration enabled us to detect signals from 12 channels in each hemisphere (i.e. a total of 24 channels from both hemispheres), which covered a 60 × 60 mm2 area of the frontal cortex. We used two different optical probes, each of which was specialized for each subject group (i.e. adults and children) and, according to the landmarks such as the glabella and the temple, we positioned the optical probe carefully so that the position was similar among all subjects. The detected signal was separated into two components that corresponded to the two wavelengths, using lock-in amplifiers. Raw optical data for the two near-infrared lights were recorded every 100 ms, simultaneously for each of the 24 channels and sent to a data collection computer via an A/D converter. The timing of each task event, such as the onset of the sample cue, was also transmitted to the data collection computer from the task-control computer by serial communication via RS232C. These digitized data were stored on the hard disk drive and were eventually transferred to magneto-optical diskettes for further off-line analysis.

Although the spatial resolution of OT differs according to the definition, the spatial accuracy of OT has been suggested to be ∼10 mm, as far as the experiment using phantom is concerned (Yamamoto et al., 2003). Concerning the safety of this technique, a previous study showed that light absorption of OT leads to tissue heating well below accepted levels of risk and concluded that the OT system is completely safe (Ito et al., 2000). Details of the OT system have also been described elsewhere (Koizumi et al., 1999, 2003).

To identify the anatomical position of each recording channel, we obtained a magnetic resonance image for two adult subjects using 1.5 T magnetic resonance imaging (MRI) (Stratis II, Premium; Hitachi Medical Corporation, Tokyo, Japan). When scanning, alfacalcidol beads were positioned at each of the 24 channels as MR markers, which can be identified on the MR image as spheres (Noguchi et al., 2002). Three-dimensional MR images with markers were reconstructed using MEDx for Linux software (Sensor systems, Sterling, VA).

Data Analysis

To analyse the raw optical data, we first defined a 20 s period after the aligned point, which was usually the onset of the sample cue, as the activation period. Second, we defined 5 s periods before and after the activation period as pre- and post-control periods, respectively. Then, to determine the background transmittance, we applied linear regression by the least squares method on transmittance during the pre- and post-control period.

After averaging over 10 trials for each condition, the relative changes in the oxyhaemoglobin concentration (Coxy) and deoxyhaemoglobin concentration (Cdeoxy) were calculated using a modified Beer–Lambert law (Maki et al., 1995). As the transmitted light detected by each detector reached the cerebral cortex after passing through the skin and the skull (McCormick et al., 1992), transmittance T(λ, t) for wavelength λ and measurement time t, is approximately:

graphic

where εoxy and εdeoxy are the molar absorption coefficients of oxy- (oxyHb) and deoxy-haemoglobin (deoxyHb), coxy and cdeoxy are the concentrations of oxyHb and deoxyHb, d is the effective path length in tissue, a(λ, t) is attenuation due to other absorption such as cytochrome aa3 except haemoglobin and sc(λ) is attenuation due to scattering in tissue. When activation occurred

graphic

where superscript ‘s’ means the value during activation. We assume a(λ, t) = as(λ, t) because the absorbance of haemoglobin is ∼10 times larger than that of cytochrome aa3, which dominates a(λ, t) mainly in the near-infrared region. By subtracting equation (1) from equation (2), we obtain

graphic

where

graphic

and

graphic

We defined

graphic

We evaluated the human brain oxygenation state by using these relative changes [Δcoxy(t), Δcdeoxy(t) and Δctotal(t)], because it is difficult to determine the effective path length d inside human tissue. To produce the topographic images, the concentration in oxyhaemoglobin detected from each channel was spatially interpolated (Maki et al., 1995).

To determine whether each recorded site was significantly activated, we compared the Coxy during the pre-control period with that during the activation period using the student’s t-test. To compare the activity level between different memory-loads, we applied a paired t-test on Coxy during the activation period excluding the first 5 s following stimulus onset (i.e. from 5 to 20 s after the sample cue).

Results

Behavioural Performance

Seven adults performed 10 trials of each load condition and 16 children performed 10 LOW trials. To examine the behavioural performance, we calculated the reaction time and accuracy (percentage correct) separately for each condition. Figure 3 summarizes the behavioural data. For the adults, the mean reaction time (RT) was significantly faster in the LOW condition than in the HIGH condition [1077 ± 376 (SD) ms for the LOW condition and 1257 ± 486 ms for the HIGH condition; P < 0.05, paired t-test], indicating that the memory-load was significantly different between the two conditions. The children’s RT was 1791 ± 419 ms, which was significantly slower than that for adults in both the HIGH and LOW conditions (P < 0.01, student’s t-test). Figure 3B shows individual data for the percentage of correct responses. All adults showed ≥80% correct responses for both conditions and the percentage correct did not differ between the two conditions (91 ± 9 (SD) % for the LOW condition and 89 ± 7% for the HIGH condition; P > 0.1, paired t-test). Four children showed relatively poor performance (<75%). The mean percentage correct for the children was 84 ± 15%, which was slightly lower than that for the adults, although the difference was not statistically significant.

Optical Topography Results in Adult Subjects

To examine the spatiotemporal characteristics of changes in Coxy, we generated dynamic topograms for each hemisphere, by interpolating the data from each channel. Figure 4 shows an example of the dynamic topogram, where changes in Coxy in the right LPFC of an adult during HIGH condition are sequentially illustrated in 2 s increments. Coxy was aligned with the onset of the sample cue. Specifically, as shown in Figure 4, Coxy gradually increased ∼5 s after the onset of the sample cue. This increase was sustained until 5–10 s after the response, and then decreased gradually to the background level.

Figure 5A shows a static topogram for the same subject represented in Figure 4, displaying the timing of peak Coxy for each hemisphere (11.2 and 10.5 s from sample cue onset for right and left hemispheres, respectively), together with the position of each channel. The activated area was maximized at one channel from each hemisphere (i.e. channels 6 and 19 for the left and right hemisphere, respectively). In the HIGH condition, all seven adults showed a significant increase in Coxy during the 20 s activation period for both channels, whereas in the LOW condition, four of the seven showed a significant increase for both channels, with two showing increases in the left hemisphere only and one in the right hemisphere only (t-test, P < 0.01; Table 1). According to the group-level analysis, both channels 6 and 19 showed significant activation in the HIGH condition (t-test, P < 0.01), although activation at only channel 6, but not channel 19, reached the significant level in the LOW condition. Thereafter, we focused on these channels (i.e. Nos 6 and 19) for further analyses of each hemisphere.

To determine the precise anatomical position of these channels, we examined an MR image of the subject whose LPFC activity has been shown in Figures 4 and 5A. Figure 5B shows a 3D MR image of the subject for each hemisphere with MR markers for each measurement point. Channels 6 and 19, which are indicated by red arrows, were located in the dorsolateral area of the PFC (BA 9/46). We confirmed that the locations of channels 6 and 19 of the other subject whose MR image were obtained were similar to those of the subject shown in Figure 5B.

As shown in the dynamic topogram in Figure 4, the change in Coxy appeared to be temporally associated with the onset of the sample cue. To examine this in detail, we generated event-triggered averaged plots of Coxy aligned at the onset of the warning signal, the sample cue and the button press. Figure 6 shows such plots of data averaged across the seven adults. This figure shows that the peak value was highest when the time was aligned with the onset of the sample cue (blue line) compared with when it was aligned with either the warning signal (red) or the button press (green). Thus, the change in Coxy appeared to be more nicely time-locked to the sample cue presentation, rather than either the warning signal or the button press, although the warning signal also appeared to provide a time-locked signal.

To test the effects of memory-load on LPFC activity, we compared Coxy in the HIGH condition with that in the LOW condition. Figure 7 shows the time course of Coxy and Cdeoxy in the HIGH and LOW conditions, in which data averaged across the seven subjects are shown. In the HIGH condition, both right and left LPFC showed significant increase in Coxy during the activation period compared with the pre-control period in the HIGH condition (P < 0.001 for both hemispheres). On the other hand, in the LOW condition, the magnitude of change in Coxy was small relative to that in HIGH condition. According to a paired t-test, the mean Coxy was significantly higher in the HIGH condition than in the LOW condition for both hemispheres (P < 0.01), indicating that the magnitude of activity in the LPFC was load dependent. For both the HIGH and LOW conditions, the activity level was not significantly different between the left and right hemispheres — although, especially in the LOW condition, the left hemisphere tended to show higher activity than the right hemisphere (paired t-test: t = 1.46, P > 0.1 for the HIGH condition; t = 2.24, 0.05 < P < 0.1 for the LOW condition). This absence of significant hemispheric difference would be due, at least in part, to the high variability in hemispheric dominance across individuals, which could also be related to probe placement or folding/localization variability across subjects.

Contrary to Coxy, Cdeoxy showed a slight decrease after the sample cue onset, although this change was not significant. This tendency for a slight decrease is consistent with many previous NIRS studies (e.g. Maki et al., 1995; Sato et al., 1999).

Optical Topography Results in Preschool Children

Figure 8A shows a dynamic topogram of the right LPFC of a preschool-age subject (6 years old) during task performance. As shown in Figure 8A, Coxy increased gradually after the onset of the sample cue. This increase was sustained until ∼5 s after the response. Then, Coxy decreased gradually to the background level. Compared with the adult subject shown in Figure 4, the overall pattern of changes in Coxy was similar in adults and children, although the magnitude of activity was larger, and active areas were more extended in children than in adults.

To examine the precise time course of changes in Coxy for this subject, we plotted averaged Coxy as a function of time triggered at the onset of sample cue for those channels that were associated with maximum activity in adults (i.e. channels 6 and 19 for the left and right hemispheres, respectively; see Fig. 5) (Fig. 8B). As shown in Figure 8B, the LPFC of both hemispheres showed a clear change in Coxy, while Cdeoxy slightly decreased. This change was characterized by a gradual increase that was time locked to the onset of the sample cue and that sustained activation until 5–10 s after the response. As summarized in Table 1, 11 of 16 subjects showed significant activation in both the right and left hemispheres, with one of the 16 and three of the 16 showing significant activation in the left or right hemisphere only, respectively (Student’s t-test, P < 0.01).

To examine the time course of changes in Coxy across our population of preschool children, we plotted mean Coxy as a function of time at channels 9 and 16 for the left and right hemispheres, respectively, across all 16 children (Fig. 9A). Similarly to individual data from a typical subject (Fig. 8B), the group mean data (Fig. 9A) also showed a significant increase in Coxy, but not in Cdeoxy, for both hemispheres, which was time locked to the onset of the sample cue. Again, these characteristics of LPFC activity during the task were similar to those of the adults. However, unlike adults, children tended to show right-hemisphere dominance, according to the same analysis used for adult subjects (paired t-test: t = 2.08, 0.05 < P < 0.1).

Since, as shown in behavioural results in Figure 3B, performance in some children were poor, we also plotted the histogram of mean Coxy after excluding the data from subjects showing poor performance (i.e. <75%; Fig. 9B). However, there was no clear difference after excluding the data, which may be because incorrect response was primarily a kind of careless miss despite remembering the sample cue locations, as all subjects extensively practiced the task to be high performance (>80%) before the recording session.

Discussion

In the present study, we measured haemodynamic changes in the LPFC of adults and preschool children using an OT technique while subjects performed an item-recognition task, which requires working memory. We observed changes in Coxy in the bilateral LPFC of adults during task performance. This change was characterized by a sustained increase that was temporally associated with the onset of the sample cue. The magnitude of activity differed between different memory-load conditions; the LPFC was more active in the high-memory-load than in the low-memory-load condition. In preschool children, we also observed haemodynamic changes in the LPFC during performance of the task. The overall pattern of these changes was similar to that of the adults, in terms of (i) a temporal association between the change in Coxy and the sample cue presentation, (ii) a sustained increase in Coxy and (iii) a decrease to the background level 5–10 s after the response. For both adults and children, the change in Coxy was accompanied by a nonsignificant trend of decrease in Cdeoxy, which has also been reported in many previous studies that used the NIRS technique (e.g. Maki et al., 1995; Sato et al., 1999). These results indicate that the LPFC is active in both preschool children and adults during working memory, suggesting that the LPFC has already developed functionally in 5–6-year-olds to enable processing of this cognitive function.

LPFC Activation Associated with Working Memory Measured by Event-related OT

The OT technique adopted here is a type of near-infrared spectroscopy (NIRS; Jöbsis, 1977), but it has an advantage over other NIRS measures: OT can measure temporal changes in haemoglobin oxygenation from multiple regions simultaneously, via multiple pairs of light emitters and detectors (Koizumi et al., 1999, 2003), whereas other NIRS techniques measure spectroscopic reflection and scattering via a single emitter-detector (Chance et al., 1993; Hoshi and Tamura, 1993; Kato et al., 1993; Villringer et al., 1993). Although OT also has some disadvantages compared with other methods (e.g. it can not measure the absolute value of each haemoglobin), it can potentially be a useful tool for functional mapping. Indeed, previous studies using OT have reported haemodynamic changes associated with various cortical functions such as motor action (Maki et al., 1995, 1996) and language processing (Watanabe et al., 1998; Sato et al., 1999; Kennan et al., 2002). These studies have adopted a blocked-task paradigm that measures accumulated activity over many trials in close succession. In contrast, Noguchi et al. (2002) recently demonstrated that an event-related design is useful in OT for studying higher cognitive functions. In the present study, we were the first to adopt such an event-related design for the study of function mediated by dorsolateral areas of the PFC and we detected event-related changes in Coxy in the dorsolateral PFC.

Our behavioural task was a type of item-recognition task, which has been extensively used to study working memory processes, including functional activation studies with PET/fMRI (e.g. Jonides et al., 1993, 1997; Belger et al., 1998; Leung et al., 2002; for reviews, see Smith and Jonides, 1999; D’Esposito et al., 2000). While the subjects performed this task, we observed activation in the dorsolateral areas of the PFC (BA 9/46). This area of activation is consistent with those reported in many previous studies (e.g. Funahashi et al., 1989; Jonides et al., 1993; Belger et al., 1998; Owen et al., 1999; Zarahn et al., 1999). Furthermore, the time course of haemodynamic changes appears to be similar to those reported in previous event-related fMRI studies (e.g. Zarahn et al., 1999; Rypma and D’Esposito, 2000; Leung et al., 2002). In addition, our task had two different memory-load conditions, which was confirmed by the differential behavioural data, and we observed load dependency in LPFC activation across different memory-load conditions. Such a load dependency in LPFC activation has also been reported in many previous studies that used PET/fMRI to examine working memory (Braver et al., 1997; Cohen et al., 1997; Jonides et al., 1997; Manoach et al., 1997; Rypma and D’Esposito, 1999). From the precise time course of changes in Coxy, both the sample cue onset and the warning signal appeared to provide a time-locked signal. This may also be consistent with previous studies, because the LPFC has been suggested to play a role in task preparation (MacDonald et al., 2000; Brass and von Cramon, 2002). Taken together, these data indicate that the changes in Coxy observed here are associated with working memory.

Thus, event-related OT is useful for examining higher order cognitive functions that are mediated by the LPFC, although this method has some disadvantages — for example, relatively low spatial resolution and the restriction of measurements to the lateral surface of the cerebral cortex. Nonetheless, OT will become a new approach to the mapping of cognitive functions, since the method has some clear advantages, in particular its complete non-invasiveness and unobtrusiveness, which makes it usable with children.

LPFC Activation in Preschool Children and Possible Application of OT on Functional Developmental Study

Since OT could measure LPFC activity associated with working memory in normal adults, we applied the same procedure to preschool children aged from 5 to 6 years. We detected significant changes in Coxy in the LPFC during the task performance. These changes were characterized by (i) a temporal association between the change in Coxy and the sample cue presentation, (ii) a sustained increase in Coxy and (iii) a decrease to the background level 5–10 s after the response. These characteristics were similar to those of adults. Therefore, it has been suggested that these changes in preschool children are associated with working memory. Thus, our findings are the first to suggest that the LPFC of 5- to 6-year-olds has developed functionally to enable processing of working memory and should make an important contribution to the understanding of the functional and/or structural development of the LPFC.

Previous psychological and morphological studies have suggested that a PFC-mediated working memory system has already developed by preschool age. For example, at the behavioural level, it was suggested that visuospatial working memory emerges at around the age of 4 years and improves between the ages of 5 and 7 years (Luciana and Nelson, 1998). With respect to brain structure, morphological studies have shown that the volume of white matter increases monotonically from the age of 4 years to young adulthood and that the volume of the frontal grey matter increases from childhood to adolescence (Giedd et al., 1999). Furthermore, the number of synapses in the PFC increases from birth to around 7 years of age (Huttenlocher, 1979). The present findings are consistent with these studies and extend their results, by providing direct evidence that development of the LPFC in preschool children (aged from 5 to 6 years) already enables processing of working memory.

Recently, Kwon et al. (2002) showed an age-related increase in LPFC activity during a working memory task in subjects aged between 7 and 22 years. Together with their findings, we suggest that functional development of the LPFC for working memory has already commenced by the age of 5 years. However, in the present study, we could not examine age-related changes, because the age of our subjects did not vary enough to investigate the correlation between age and activity. Further studies using OT in children across a wider age range (e.g. from 3 to 6 years of age) will contribute to the understanding of age-related functional maturation over a range of preschool ages.

Comparison between Adults and Preschool Children

In the present study, according to detailed characteristics of the haemodynamic changes in adults and preschool children, the LPFC of children appeared to be more highly and broadly active than that of adults. This may reflect the developmental change of LPFC function to enable more efficient processing of such an important cognitive function as working memory (Casey et al., 1997; Tamm et al., 2002). However, the difference in magnitude and extension may simply reflect some structural differences between adults and children. Indeed, we could not directly compare the difference in activity between adults and children, owing to differences in brain structure, skull, hairline and so on, which can cause differences in light absorption between adults and children. This would also make it difficult to address whether our results are consistent with the morphological studies that have found protracted increase in white matter volume of the PFC from the age of 8 years to young adulthood (Giedd et al., 1999).

Furthermore, an inverse tendency was observed with respect to hemispheric dominance in adults and children: the left LPFC tended to be dominant in adult subjects, especially in the LOW condition, while the right LPFC tended to be dominant in children. This difference might reflect developmental changes in functional specialization of the LPFC for working memory, as reported in previous event-related potential studies of the development of language systems (Mills et al., 1997; Neville and Bavelier, 1998). However, the hemispheric differences observed here were not very significant (i.e. 0.05 < P < 0.1) and it is possible that adults and children might adopt different strategies for performing the present task. Therefore, these problems regarding hemispheric difference remain to be solved by further studies that use a more appropriate method.

Conclusions

We first demonstrated that OT with an event-related paradigm can examine the neural correlate of working memory mediated by the LPFC. This can introduce a new dimension to the mapping of LPFC function in normal children as well as in adults. In addition, for the first time, we demonstrated that the LPFC of preschool children is active during the operation of working memory. Although we could not comprehensively compare the activity of adults and children, or in children at different stages of development, our results extend the understanding of not only the functional development, but also the neural basis of cognitive functions mediated by the LPFC.

The authors are grateful to G. Hirano and T. Tanase for their technical and administrative assistance and encouragement throughout this study. The authors also thank Dr T. Toyoshima for his helpful comments on ethics and M. Ishimado and other kindergarten teachers at Shiratori Kindergarten for treatment of children. This work was supported in part by Grant-in-Aid for JSPS Fellows (15009927) from the Japan Society for the Promotion of Science to S.T. and by Grant-in-Aid for Exploratory Research (15650124) from the Ministry of Education, Culture, Sports, Science and Technology to T.S.

Figure 1. Schematic drawings of the item-recognition task with two memory-load conditions. The subjects remembered the location of two or four sample cues in the LOW and HIGH conditions, respectively. Then they were required to report whether the location of the test cue was identical to one of the sample cues.

Figure 1. Schematic drawings of the item-recognition task with two memory-load conditions. The subjects remembered the location of two or four sample cues in the LOW and HIGH conditions, respectively. Then they were required to report whether the location of the test cue was identical to one of the sample cues.

Figure 2. Configuration of emitters, detectors and measurement points for a hemisphere.

Figure 2. Configuration of emitters, detectors and measurement points for a hemisphere.

Figure 3. Behavioural results. (A) Mean reaction times are shown separately by condition. (B) Percentage correct for each individual subject for each condition. Filled quarries represent the data from each subject and open circles show mean percentage correct for each condition. Error bars indicate SE. *P < 0.05.

Figure 3. Behavioural results. (A) Mean reaction times are shown separately by condition. (B) Percentage correct for each individual subject for each condition. Filled quarries represent the data from each subject and open circles show mean percentage correct for each condition. Error bars indicate SE. *P < 0.05.

Figure 4. Dynamic topogram of changes in oxyhaemoglobin concentration in the right LPFC of an adult during a working memory task under the HIGH condition. Averaged data over 10 trials for a typical subject are shown for every 2 s, from –6 s to 24 s. The onset of the sample cue is represented by 0 s. The approximate recording area is indicated as a shaded area of the frontal cortex in the lower right illustration of this figure.

Figure 4. Dynamic topogram of changes in oxyhaemoglobin concentration in the right LPFC of an adult during a working memory task under the HIGH condition. Averaged data over 10 trials for a typical subject are shown for every 2 s, from –6 s to 24 s. The onset of the sample cue is represented by 0 s. The approximate recording area is indicated as a shaded area of the frontal cortex in the lower right illustration of this figure.

Figure 5. Static topogram and the anatomical sites of focused channels. (A) Static topograms at the peak time for each hemisphere are shown, together with the point of each of the 24 channels. The data from the same subject as in Figure 4. Blue arrows show the channels that showed highest activity in each hemisphere. (B) 3D MR image of the subject shown in (A). Blue dots indicate each recording channel shown. Channels indicated by a red arrow correspond to channels 6 and 19 in (A), for left and right hemispheres, respectively.

Figure 5. Static topogram and the anatomical sites of focused channels. (A) Static topograms at the peak time for each hemisphere are shown, together with the point of each of the 24 channels. The data from the same subject as in Figure 4. Blue arrows show the channels that showed highest activity in each hemisphere. (B) 3D MR image of the subject shown in (A). Blue dots indicate each recording channel shown. Channels indicated by a red arrow correspond to channels 6 and 19 in (A), for left and right hemispheres, respectively.

Figure 6. Temporal change in the oxyhaemoglobin (oxy-Hb) concentration across different task events. (A) and (B) show the left and right hemisphere, respectively. Data for the group mean (n = 7) in the HIGH condition are shown. The blue line shows the time course when aligned at the onset of the sample cue. The red and green lines show the time course when aligned at the warning signal and the button press, respectively. In each hemisphere, changes in oxy-Hb were temporally associated with the sample cue presentation.

Figure 6. Temporal change in the oxyhaemoglobin (oxy-Hb) concentration across different task events. (A) and (B) show the left and right hemisphere, respectively. Data for the group mean (n = 7) in the HIGH condition are shown. The blue line shows the time course when aligned at the onset of the sample cue. The red and green lines show the time course when aligned at the warning signal and the button press, respectively. In each hemisphere, changes in oxy-Hb were temporally associated with the sample cue presentation.

Figure 7. Load-dependent difference in changes in oxyHb and deoxyHb. (A) and (B) show the left and right hemispheres, respectively. Time courses of changes in Coxy and Cdeoxy for seven subjects are shown separately by HIGH and LOW conditions. The green and orange lines show the HIGH and LOW conditions, respectively, and the solid and dotted lines show Coxy and Cdeoxy, respectively. Shading indicates standard errors.

Figure 7. Load-dependent difference in changes in oxyHb and deoxyHb. (A) and (B) show the left and right hemispheres, respectively. Time courses of changes in Coxy and Cdeoxy for seven subjects are shown separately by HIGH and LOW conditions. The green and orange lines show the HIGH and LOW conditions, respectively, and the solid and dotted lines show Coxy and Cdeoxy, respectively. Shading indicates standard errors.

Figure 8. (A) Dynamic topogram of changes in the concentration of oxyhaemoglobin in the right LPFC of a 6-year-old child. The format is the same as that in Figure 3. (B) Time course of changes in oxy- and deoxyhaemoglobin concentrations in the LPFC of the child are shown in (A). Solid and dotted lines show oxy- and deoxyhaemoglobin, respectively. Data from channels 6 and 19 are shown for the left and right hemispheres, respectively.

Figure 8. (A) Dynamic topogram of changes in the concentration of oxyhaemoglobin in the right LPFC of a 6-year-old child. The format is the same as that in Figure 3. (B) Time course of changes in oxy- and deoxyhaemoglobin concentrations in the LPFC of the child are shown in (A). Solid and dotted lines show oxy- and deoxyhaemoglobin, respectively. Data from channels 6 and 19 are shown for the left and right hemispheres, respectively.

Figure 9. Time course of changes in oxyhaemoglobin concentration in the LPFC of children. (A) Mean concentration change across the population (n = 16) for the LPFC. (B) Data excluding subjects with poor performance (<75%) are shown (n = 12). Solid and dotted lines show oxy- and deoxyhaemoglobin, respectively. Shading indicates standard errors.

Figure 9. Time course of changes in oxyhaemoglobin concentration in the LPFC of children. (A) Mean concentration change across the population (n = 16) for the LPFC. (B) Data excluding subjects with poor performance (<75%) are shown (n = 12). Solid and dotted lines show oxy- and deoxyhaemoglobin, respectively. Shading indicates standard errors.

Table 1


 The number of subjects showing a significant increase in Coxy

 R only L only Both Total 
Adults HIGH (n = 7)  7  7 
Adults LOW (n = 7)  4  7 
Children (n = 16) 11 15 
 R only L only Both Total 
Adults HIGH (n = 7)  7  7 
Adults LOW (n = 7)  4  7 
Children (n = 16) 11 15 

R and L indicate right and left hemispheres, respectively.

References

Baddeley A (
1986
) Working memory. Oxford: Oxford University Press.
Baddeley A (
1992
) Working memory.
Science
 
255
:
556
–559.
Bauer RH, Fuster JM (
1976
) Delayed-matching and delayed-response deficit from cooling dorsolateral prefrontal cortex in monkeys.
J Comp Physiol Psychol
 
90
:
293
–302.
Belger A, Puce A, Krystal JH, Gore JC, Goldman-Rakic PS, McCarthy G (
1998
) Dissociation of mnemonic and perceptual processes during spatial and nonspatial working memory using fMRI.
Hum Brain Mapp
 
6
:
14
–32.
Brass M, von Cramon DY (
2002
) The role of the frontal cortex in task preparation.
Cereb Cortex
 
12
:
908
–914.
Braver TS, Cohen JD, Nystrom LE, Jonides J, Smith EE, Noll DC (
1997
) A parametric study of prefrontal cortex involvement in human working memory.
Neuroimage
 
5
:
49
–62.
Casey BJ, Cohen JD, Jezzard P, Turner R, Noll DC, Trainor RJ, Giedd J, Kaysen D, Hertz-Pannier L, Rapoport JL (
1995
) Activation of prefrontal cortex in children during a nonspatial working memory task with functional MRI.,
Neuroimage
 
2
:
221
–229.
Casey BJ, Trainor RJ, Orendi JL, Schubert AB, Nystrom LE, Giedd JN, Castellanos FX, Haxby JV, Noll DC, Cohen JD, Forman SD, Dahl RE, Rapoport JL (
1997
) A developmental functional MRI study of prefrontal activation during performance of a go–no-go task.
J Cogn Neurosci
 
9
:
835
–847.
Casey BJ, Giedd JN, Thomas KM (
2000
) Structural and functional brain development and its relation to cognitive development.
Biol Psychol
 
54
:
241
–257.
Chance B, Zhuang Z, Unah C, Alter C, Lipton L (
1993
) Cognition-activated low-frequency modulation of light absorption in human brain.
Proc Natl Acad Sci USA
 
90
:
3770
–3774.
Cohen JD, Perlstein WM, Braver TS, Nystrom LE, Noll DC, Jonides J, Smith EE (
1997
) Temporal dynamics of brain activation during a working memory task.
Nature
 
386
:
604
–608.
Courtney SM, Petit L, Maisog J, Ungerleider LG, Haxby, V (
1998
) An area specialized for spatial working memory in human frontal cortex.
Science
 
279
:
1347
–1350.
D’Esposito M, Postle BR, Rypma B (
2000
) Prefrontal cortical contributions to working memory: evidence from event-related fMRI studies.
Exp Brain Res
 
133
:
3
–11.
Freedman M, Oscar-Berman M (
1986
) Bilateral frontal lobe disease and selective delayed response deficits in humans.
Behav Neurosci
 
100
:
337
–342.
Funahashi S, Bruce CJ, Goldman-Rakic PS (
1993
) Dorsolateral prefrontal lesions and oculomotor delayed-response performance: evidence for mnemonic scotomas.
J Neurosci
 
13
:
1479
–1497.
Funahashi S, Kubota K (
1994
) Working memory and prefrontal cortex.
Neurosci Res
 
21
:
1
–11.
Funahashi S, Bruce CJ, Goldman-Rakic PS (
1989
) Mnemonic coding of visual space in primate prefrontal neurons revealed by oculomotor paradigms.
J Neurophysiol
 
63
:
814
–831.
Fuster JM (
1997
) The prefrontal cortex. Philadelphia, PA: Lippincott-Raven.
Giedd JN, Blumenthal J, Jeffries NO, Castellanos FX, Liu H, Zijdenbos A, Paus T, Evans AC, Rapoport JL (
1999
) Brain development during childhood and adolescence: a longitudinal MRI study.
Nat Neurosci
 
2
:
861
–863.
Goldman PS, Rosvold E (
1970
) Localization of function within the dorsolateral PFC of the rhesus monkey.
Exp Neurol
 
27
:
291
–304.
Goldman-Rakic PS (
1995
) Cellular basis of working memory.
Neuron
 
14
:
477
–485.
Hoshi Y, Tamura M (
1993
) Detection of dynamic changes in cerebral oxygenation coupled to neuronal function during mental work in man.
Neurosci Lett
 
150
:
5
–8.
Huttenlocher PR (
1979
) Synaptic density in human frontal cortex — developmental changes and effects of aging.
Brain Res
 
163
:
195
–205.
Huttenlocher PR (
2003
) Basic neuroscience research has important implications for child development.
Nat Neurosci
 
6
:
541
.
Ito Y, Kennan RP, Watanabe E, Koizumi H (
2000
) Assessment of heating effects in skin during continuous wave near infrared spectroscopy.
J Biom Opt
 
5
:
383
–390.
Jöbsis FF (
1977
) Noninvasive infrared monitoring of cerebral and myocardial oxygen sufficiency and circulatory parameters.
Science
 
198
:
1264
–1267.
Jonides J, Smith EE, Koeppe RA, Awh E, Minoshima S, Mintun MA (
1993
) Spatial working memory in humans as revealed by PET.
Nature
 
363
:
623
–625.
Jonides J, Schumacher EH, Smith EE, Lauber EJ, Awh E, Minoshima S, Koeppe RA (
1997
) Verbal working memory load affects regional brain activation as measured by PET.
J Cogn Neurosci
 
9
:
462
–475.
Kail R, Salthouse TA (
1994
) Processing speed as a mental capacity.
Acta Psychol
 
86
:
199
–225.
Kato T, Kamei A, Takashima S, Ozaki T (
1993
) Human visual cortical function during photic stimulation monitoring by means of near-infrared spectroscopy.
J Cereb Blood Flow Metab
 
13
:
516
–520.
Kennan RP, Kim D, Maki A, Koizumi H, Constable RT (
2002
) Non-invasive assessment of language lateralization by transcranial near infrared optical topography and functional MRI.,
Hum Brain Mapp
 
16
:
183
–189.
Koizumi H, Yamashita A, Maki A, Yamamoto T, Ito Y, Itagaki H, Kennan R (
1999
) Higher-order brain function analysis by trans-cranial dynamic near-infrared spectroscopy imaging.
J Biomed Opt
 
94
:
403
–413.
Koizumi H, Yamamoto T, Maki A, Yamashita Y, Sato H, Kawaguchi H, Ichikawa N (
2003
) Optical topography: practical problems and new applications.
Appl Phys
 
42
:
3054
–3062.
Kwon H, Reiss AL, Menon V (
2002
) Neural basis of protracted developmental changes in visuo-spatial working memory.
Proc Natl Acad Sci USA
 
99
:
13336
–13341.
Leung HC, Gore JC, Goldman-Rakic PS (
2002
) Sustained mnemonic response in the human middle frontal gyrus during on-line storage of spatial memoranda.
J Cogn Neurosci
 
14
:
659
–671.
Luciana M, Nelson CA (
1998
) The functional emergence of prefrontally-guided working memory systems in four- to eight-year-old children.
Neuropsychologia
 
36
:
273
–293.
McCarthy G, Blamire AM, Puce A, Nobre AC, Bloch G, Hyder F, Goldman-Rakic PS, Shulman RG (
1994
) Functional magnetic resonance imaging of human prefrontal cortex during a spatial working memory task.
Proc Natl Acad Sci USA
 
91
:
8690
–8694.
McCormick PW, Stewart M, Lewis G, Dujovny M, Ausman JI (
1992
) Intracerebral penetration of infrared light.
J Neurosurg
 
76
:
315
–318.
MacDonald AW III, Cohen JD, Stenger VA, Carter CS (
2000
) Dissociating the role of the dorsolateral prefrontal and anterior cingulate cortex in cognitive control.
Science
 
288
:
1835
–1838.
Maki A, Yamashita Y, Ito Y, Watanabe E, Mayanagi Y, Koizumi H (
1995
) Spatial and temporal analysis of human motor activity using non-invasive NIR topography.
Med Phys
 
22
:
1997
–2005.
Maki A, Yamashita Y, Watanabe E, Koizumi H (
1996
) Visualizing human motor activity by using non-invasive optical topography.
Frontiers Med Biol Eng
 
7
:
285
–297.
Manoach DS, Schlaug G, Siewert B, Darby DG, Bly BM, Benfield A, Edelman RR, Warach S (
1997
) Prefrontal cortex fMRI signal changes are correlated with working memory load.
Neuroreport
 
8
:
545
–549.
Mills D, Coffey-Corina S, Neville HJ (
1997
) Language comprehension and cerebral specialization from 13 to 20 months.
Dev Neuropsychol
 
13
:
397
–445.
Nelson CA, Monk CS, Lin J, Carver LJ, Thomas KM, Truwit CL (
2000
) Functional neuroanatomy of spatial working memory in children.
Dev Psychol
 
36
:
109
–116.
Neville HJ, Bavelier D (
1998
) Neural organization and plasticity of language.
Curr Opin Neurobiol
 
8
:
254
–258.
Noguchi Y, Takeuchi T, Sakai KL (
2002
) Lateralized activation in the inferior frontal cortex during syntactic processing: event-related optical topography study.
Hum Brain Mapp
 
17
:
89
–99.
Owen AM, Evans AC, Petrides M (
1996
) Evidence for a two-stage model of spatial working memory processing within the lateral frontal cortex: a positron emission tomography study.
Cereb Cortex
 
6
:
31
–38.
Owen AM, Herrod NJ, Menon DK, Clark JC, Downey SPMJ, Carpenter TA, Minhas PS, Turkheimer FE, Williams EJ, Robbins TW, Sahakian BJ, Petrides M, Pickard JD (
1999
) Redefining the functional organization of working memory processes within human lateral prefrontal cortex.
Eur J Neurosci
 
11
:
567
–574.
Rypma B, D’Esposito M (
1999
) The roles of prefrontal brain regions in components of working memory: effects of memory load and individual differences.
Proc Natl Acad Sci USA
 
96
:
6558
–6563.
Rypma B, D’Esposito M (
2000
) Isolating the neural mechanisms of age-related changes in human working memory.
Nat Neurosci
 
3
:
509
–515.
Sato H, Takeuchi T, Sakai KL (
1999
) Temporal cortex activation during speech recognition: an optical topography study.
Cognition
 
73
:B
55
–B66.
Sawaguchi T, Iba M (
2001
) Prefrontal cortical representation of visuospatial working memory in monkeys examined by local inactivation with muscimol.
J Neurophysiol
 
86
:
2041
–2053.
Sawaguchi T, Yamane I (
1999
) Properties of delay-period neuronal activity in the monkey dorsolateral prefrontal cortex during a spatial delayed matching-to-sample task.
J Neurophysiol
 
82
:
2070
–2080.
Shimamura AP (
1994
) Memory and frontal lobe function. In: The cognitive neurosciences (Gazzaniga MS, ed.), pp.
803
–815. Cambridge, MA: MIT Press.
Smith EE, Jonides J (
1999
) Storage and executive processes in the frontal lobes.
Science
 
283
:
1657
–1661.
Sweeney JA, Mintun MA, Kwee S, Wiseman MB, Brown DL, Rosenberg DR, Carl JR (
1996
) Positron emission tomography study of voluntary saccadic eye movements and spatial working memory.
J Neurophysiol
 
75
:
454
–468.
Tamm L, Menon V, Reiss AL (
2002
) Maturation of brain function associated with response inhibition.
J Am Acad Child Adolesc Psychiatry
 
41
:
1231
–1238.
Thomas KM, King SW, Franzen PL, Welsh TF, Berkowitz AL, Noll DC, Birmaher V, Casey BJ (
1999
) A developmental functional MRI study of spatial working memory.
Neuroimage
 
10
:
327
–338.
Villringer A, Planck J, Hock C, Schleinkofer L, Dirnagl U (
1993
) Near infrared spectroscopy (NIRS): a new tool to study hemodynamic changes during activation of brain function in human adults.
Neurosci Lett
 
154
:
101
–104.
Watanabe E, Maki A, Kawaguchi K, Takashiro Y, Yamashita H, Koizumi H (
1998
) Non-invasive assessment of language dominance with near-infrared spectroscopic mapping.
Neurosci Lett
 
256
:
49
–52.
Yamamoto T, Okada E, Kawaguchi F, Maki A, Yamada Y, Koizumi H (
2003
) Optical fiber arrangement of optical topography for spatial resolution improvement.
Proc SPIE
 
4955
:
487
–496.
Yamashita Y, Maki A, Koizumi H (
1996
) Near-infrared topographic measurement system: imaging of absorbers localized in a scattering medium.
Rev Sci Instrum
 
67
:
730
–732.
Zald DH, Iacono WG (
1998
) The development of spatial working memory abilities.
Dev Neuropsychol
 
14
:
563
–578.
Zarahn E, Aguirre GK, D’Esposito M (
1999
) Temporal isolation of the neural correlates of spatial mnemonic processing with fMRI.
Cogn Brain Res
 
7
:
255
–268.