Abstract

Children begin performing similarly to adults on tasks requiring executive functions in late childhood, a transition that is probably due to neuroanatomical fine-tuning processes, including myelination and synaptic pruning. In parallel to such structural changes in neuroanatomical organization, development of functional organization may also be associated with cognitive behaviors in children. We examined 6- to 10-year-old children's cortical thickness, functional organization, and cognitive performance. We used structural magnetic resonance imaging (MRI) to identify areas with cortical thinning, resting-state fMRI to identify functional organization in parallel to cortical development, and working memory/response inhibition tasks to assess executive functioning. We found that neuroanatomical changes in the form of cortical thinning spread over bilateral frontal, parietal, and occipital regions. These regions were engaged in 3 functional networks: sensorimotor and auditory, executive control, and default mode network. Furthermore, we found that working memory and response inhibition only associated with regional functional connectivity, but not topological organization (i.e., local and global efficiency of information transfer) of these functional networks. Interestingly, functional connections associated with “bottom-up” as opposed to “top-down” processing were more clearly related to children's performance on working memory and response inhibition, implying an important role for brain systems involved in late childhood.

Introduction

By 6 years of age, the brain volume is already ∼95% of the adult size, implying that the majority of volumetric change has already occurred. However, at this period in development the brain is undergoing a critical fine-tuning process along with forms of cellular maturation, such as myelination and synaptic pruning, allowing for cognitive development (Rakic et al. 1994; Huttenlocher and Dabholkar 1997; Amso and Casey 2006; Petanjek et al. 2008). Cortical thickness, assessed using structural magnetic resonance imaging (MRI), is often used to identify anatomical substrates in relation to improved cognitive ability in childhood. For example, studies using structural MRI demonstrate that decreases in frontal cortical thickness, which likely accompany fine-tuning processes of synaptic circuitry and gains in myelination, occur during developmental periods (Huttenlocher and Dabholkar 1997; Sowell, Thompson, Holmes et al. 1999; Sowell, Thompson, Leonard et al. 2004; Petanjek et al. 2008) that parallel increases in executive functioning (Spear 2000).

Nevertheless, executive functions emerge from the complex interaction of multiple brain areas (Fassbender et al. 2004; Jurado and Rosselli 2007) that are not apparent using structural MRI. In contrast, resting-state functional magnetic resonance imaging (rsfMRI) enables a summary of complex patterns of brain functional organization, which can be examined in relation to cognitive functioning. Several studies with adults examining functional network properties in the resting brain report that greater efficiency in information transfer in the default mode network (DMN) relates to high intelligence quotients (IQs) (Song et al. 2008; van den Heuvel et al. 2009). Moreover, amongst adults imaged in a resting state, functional connectivity associates with executive functioning and associative memory processes. For example, performance on the Trail Making Test, which measures the ability to switch between competing mental sets while performing a visuomotor search task, is associated with the connectivity among the lateral parietal nodes of the executive-control network (Seeley et al. 2007). Likewise, working memory performance is inversely related to connectivity between dorsolateral prefrontal cortex (a classic task-positive area) and the medial prefrontal cortex (a task-negative region) (Hampson et al. 2010). RsfMRI is used increasingly in studies of development. For example, rsfMRI studies demonstrate a shift from diffuse to focal activation patterns, and simultaneous pruning of local connectivity and strengthening of long-range connectivity with age (Uddin et al. 2010), as well as the emergence of functional segregation and integration in infancy (Fransson et al. 2009; Gao et al. 2009) and childhood (Fair et al. 2008). The insights gained from rsfMRI studies also shed light on potentially disrupted functional networks underlying atypical cognitive development associated with neurodevelopmental disorders (Cao et al. 2009; Church et al. 2009; Vogel et al. 2010; Bray et al. 2011). However, to date, there is very limited research with rsfMRI to investigate the complex relationships between cognitive performance and the integrity of neural networks in children.

In this study, we examined anatomical fine-tuning processes as reflected by cortical thinning that can only partially explain age-related improvement in cognitive performance in late childhood. The functional networks in parallel with cortical thinning could explain adult-like cognitive performance in children. To accomplish this goal, we first identified functional networks that contained cortical regions simultaneously undergoing structural changes with structural MRI to identify age-related changes in cortical thickness. We then identified functional networks associated with these regions using rsfMRI. Finally, we used graph and functional connectivity analyses to examine the extent to which functional integrity (i.e., efficiency of information transfer and connection strength) predicted performance on memory and response inhibition in late childhood. We focused on these specific cognitive measures because of reported developmental trends over late childhood (Hitch et al. 1983; Pickering 2001; Vuontela et al. 2003; Tillman et al. 2008; Urben et al. 2011). Moreover, brain regions that mediate these functions, such as prefrontal cortex, undergo considerable anatomical development during late childhood (Huttenlocher and Dabholkar 1997; Sowell et al. 2002; Lenroot and Giedd 2006; Muftuler et al. 2011).

Materials and Methods

Subjects

Fifty ethnically Chinese Singaporean children aged from 6 to 10 years old [mean = 8.05 years, standard deviation (SD) = 1.43 years] were recruited from an existing population-based cohort study of myopia (Dirani et al. 2010). During recruitment we excluded boys with an existing diagnosis of chronic medical conditions (e.g., cancer, congenital abnormalities) and/or mental disorders (e.g., attention-deficit hyperactivity disorder (ADHD), Autism). Written consent was obtained from participants' parents under the approval of the Institutional Review Board of the National University of Singapore.

Cognitive Tasks

The Cambridge Neuropsychological Test of Automated Battery (CANTAB) includes language-independent cognitive tests (Luciana and Nelson 2002) administered on a computer fitted with a touch-sensitive screen and 2-button response pad. Participants were first screened on 2 motor and learning tasks to verify the ability to follow simple instructions. Subsequently, participants performed the following tasks: spatial working memory (SWM), delayed-matching-to-sample task (DMS), and the stop-signal task (SST).

Spatial Working Memory

The SWM is a self-ordered searching task that requires participants to maintain and update spatial information in working memory. On each trial, participants are asked to find blue tokens hidden in 3, 4, or 6 boxes. They are told that the tokens are hidden one at a time and that over the course of a trial they are never hidden in the same box twice. Within each trial, a token will ultimately be hidden in each box. Thus, participants need to make multiple searches within 1 trial: as soon as a token is found, a new search begins. Participants make “within search errors,” when they return to the same empty box before finding a token. They commit “between search errors” when they return to a box where a token has been found in a prior search. We considered only the trials with 6 boxes. We combined these 2 error types and computed a total error score to measure children's ability to keep spatial information active in working memory.

Delayed Matching to Sample

The DMS is a test of visual working memory (Miller et al. 1996; Swartz et al. 1996). On each trial a complex visual pattern (the sample) is briefly shown. The sample is then covered and participants see 4 patterns below the sample after a delay of 0, 4, or 12 s. Subjects are told to select the pattern identical to the sample. A DMS total correct score is reported as the number of trials for which the participant selects the correct stimulus in trials of all delays and used to measure children's ability to keep visual information active in working memory over a short period of time.

Stop-Signal Task

The SST is a classic stop-signal response inhibition test. During the SST, go-stimuli are displayed for 1000 ms as left and right arrows positioned in the center of a computer screen. The left and right arrows appear with equal frequency throughout the task. The subjects are told to press the left button on the response pad when a left-pointing arrow appears on the screen and the right button when a right-pointing arrow appears. An auditory tone (i.e., the “stop-stimulus”) is generated by the computer and presented randomly on 25% of the trials. The subjects are instructed to withhold responding on trials accompanied by the auditory tone. Stop-signal delays are initially set at 250 ms and dynamically adjusted ±50 ms contingent on children's performance on the previous trial. Successfully inhibited stop-trials are followed by a 50 ms increase in stop-signal delay, and unsuccessfully inhibited stop-trials are followed by a 50 ms decrease in stop-signal delay. The task is designed to approximate successful inhibition on 50% of the stop-trials. The stop signal reaction time (SSRT) reflects the speed of the stop process relative to the go process, and is estimated by subtracting the time interval between the go-and-stop stimuli from the time required to process and perform a controlled motor response. It has widely been adopted as a measure of impulse control (Alderson et al. 2007).

MRI Data Acquisition

MRI was performed on a 3 T Siemens Magnetom Trio Tim scanner using a 32-channel head coil at Clinical Imaging Research Centre of the National University of Singapore. The image protocols were: (i) high-resolution isotropic T1-weighted Magnetization Prepared Rapid Gradient Recalled Echo (MPRAGE; 190 slices, 1 mm thickness, in-plane resolution 1 mm, no inter-slice gap, sagittal acquisition, field of view 190 × 190 mm, matrix = 190 × 190, repetition time = 2000 ms, echo time = 2.08 ms, inversion time = 850 ms, flip angle = 9°); (ii) isotropic axial rsfMRI imaging protocol (single-shot echo-planar imaging; 42 slices with 3 mm slice thickness, no inter-slice gaps, matrix = 64 × 64, field of view = 190 × 190 mm, repetition time = 2400 ms, echo time = 27 ms, flip angle = 90°, scanning time = 6 min). The children were asked to close their eyes during the rsfMRI scan.

The image quality was verified immediately after the acquisition through visual inspection when children were still in the scanner. If the motion artifact was large, a repeated scan was conducted. The image was removed from the study if no acceptable image was acquired after 3 repetitions.

Cortical Thickness Analysis

A Markov random field (MRF) model was used to label each voxel in the T1-weighted image as gray matter (GM), white matter (WM), or CSF (Fischl et al. 2002). In the MRF model, the prior probability of each structure was computed based on the manual segmentation of 30 subjects randomly selected from our sample. A cortical inner surface was constructed at the boundary between GM and WM and then propagated to its outer surface at the boundary between GM and CSF. The cortical thickness was measured as the distance between the corresponding vertices on the inner and outer surfaces (Fischl and Dale 2000) and represented on the inner surface. The cortical surface (the middle surface obtained by averaging the inner and outer cortical surfaces) of individuals was mapped to that of an atlas using multi-manifold large deformation diffeomorphic metric mapping (MM-LDDMM) (Zhong and Qiu 2010) that simultaneously aligns the cortical surface and 9 gyral curves (precentral gyrus, postcentral gyrus, superior temporal gyrus, middle temporal gyrus, lingual gyrus, cuneus gyrus, anterior border of cuneus gyrus, posterior border of precuneus, and paracentral gyrus). The gyral curves were semi-automatically delineated on the cortical surface using dynamic programming. After the surface registration, individual cortical thickness data were transferred to the cortical surface of the atlas based on the surface correspondence derived from the MM-LDDMM.

RsfMRI Preprocessing

The rsfMRI data were first processed with slice timing, motion correction, skull stripping, band-pass filtering (0.01–0.08 Hz), and grand mean scaling of the data to whole-brain-modal value of 100. To quantify the quality of rsfMRI data in terms of head motion, displacement due to motion averaged over the image volume was calculated for individual subjects. Its mean and standard deviation were, respectively, 0.05 and 0.04 mm. Then, the rsfMRI signals due to effects of nuisance variables, including 6 parameters obtained by motion correction, ventricular and white matter signals, were removed. Subsequently, the fMRI data were represented on the cortical surface (Qiu et al. 2006) and transformed to the cortical surface of the atlas based on the MM-LDDMM transformation discussed above.

Statistical Analysis

Cortical Thickness

We examined the association of cortical thickness with age. The thickness data on the atlas cortical surface were smoothed using a 20-mm full-width-at-half-maximum Gaussian filter (Chung et al. 2005). Linear regression with age as a main factor was examined at each vertex. Results at each surface vertex were thresholded at the level of significance (P < 0.005) and then corrected for multiple comparisons at the cluster level of significance (P < 0.05). Each cluster size must be >502 mm2, which was determined based on random field theory (Chung et al. 2010).

Functional Network Analysis

We first performed a seed-based correlation analysis to examine which cortical regions are functionally connected to those anatomically developing regions in late childhood. The seed regions were chosen at locations with the peak t-values of each cluster with significant age effects on cortical thickness from the analysis described above. The Pearson's correlation coefficients with the seed regions were computed for each vertex on the atlas cortical surface and then transformed to “z-value” using Fisher's r-to-z transformation (Zar 1996). The second-level fMRI analysis was finally performed to identify significant connections between the seed regions with the rest of the cortex via 1-sample Student's t-test. The results at each surface vertex were thresholded at the level of significance (P < 0.005) and then corrected for multiple comparisons at the cluster level of significance (P < 0.05). Each cluster size was required to be >129 mm2, which was determined based on random field theory (Chung et al. 2010).

We subsequently identified functional networks in parallel with cortical thickness development in late childhood. Based on the functional connectivity maps obtained from the above seed-based analysis, the cortex was parcellated into 30 regions (15 regions per hemisphere, Fig. 3). We considered these regions as nodes in the following functional network analysis. The functional organization was represented for individual subjects using a correlation matrix with its element as Pearson's correlation coefficients of the fMRI signals between any 2 cortical regions. The elements of this matrix with significant positive correlation (P < 0.005) were retained after correcting for multiple comparisons using false discovery rate. The functional networks were identified using graph modularity analysis on the average correlation matrix among all the subjects such that the strength of the connections among the cortical regions within the same network was high compared with that across 2 networks. We then computed graph properties, such as characteristic path length and mean clustering coefficient (Latora and Marchiori 2001; Humphries and Gurney 2008) for each network at the level of the individual subject. The characteristic path length as the average shortest path on the network was used to measure how efficiently information can travel between distant nodes. The mean clustering coefficient was computed to measure how richly integrated the local environment was. These 2 measures indicated the efficiency of the network to exchange information at the global level and the extent of local inter-connectivity of a network (Latora and Marchiori 2001).

Functional Networks in Relation to Executive Functions

We first tested whether any functional network identified using the aforementioned graph modularity analysis was associated with executive functions in late childhood. Associations were examined using Pearson's correlation analysis between cognitive measures and the global topology (mean clustering coefficient and characteristic path length) of each functional network. We also examined the association between functional connectivity and executive functions. Linear regression was used with the significant functional connectivity strength obtained from the aforementioned seed-based analysis as a main factor and with the measure of executive functions as a dependent variable. Results at each surface vertex were thresholded at the level of significance (P < 0.005) and then corrected for multiple comparisons at the cluster level of significance (P < 0.05). Each cluster size must be >129 mm2 determined based on random field theory (Chung et al. 2010).

Results

Cortical Thickness and Executive Functions in Late Childhood

Age-related cortical thinning in late childhood was most apparent in the frontal, parietal, and occipital regions, including bilateral ventral medial prefrontal cortices (vmPFC), right inferior frontal gyrus (R-IFG), right intraparietal sulcus and lateral occipital cortex (R-IPS/LOC), left superior and inferior parietal lobules (L-SPL, L-IPL), and left lateral occipital cortex (L-LOC) and left cuneus (L-Cun) (Fig. 1).

Figure 1.

Statistical maps of cortical thinning in late childhood. Vertexwise analyses were corrected for multiple comparisons (t-value = −2.68, P < 0.05 corrected).L-IPL/LOC, left inferior parietal lobule/lateral occipital cortex; L-Cun/SPL, left cuneus/superior parietal lobule; L-vmPFC, left ventral medial prefrontal cortex; R-IPS/LOC, right intraparietal sulcus/lateral occipital cortex; R-IFG, right inferior frontal gyrus; R-vmPFC, right ventral medial prefrontal cortex.

Figure 1.

Statistical maps of cortical thinning in late childhood. Vertexwise analyses were corrected for multiple comparisons (t-value = −2.68, P < 0.05 corrected).L-IPL/LOC, left inferior parietal lobule/lateral occipital cortex; L-Cun/SPL, left cuneus/superior parietal lobule; L-vmPFC, left ventral medial prefrontal cortex; R-IPS/LOC, right intraparietal sulcus/lateral occipital cortex; R-IFG, right inferior frontal gyrus; R-vmPFC, right ventral medial prefrontal cortex.

Performance on spatial (total errors of the SWM task, r = −0.59, P < 0.001) and visual working (total correct score in the DMS task, r = 0.45, P = 0.001) memory tasks was improved with age. Moreover, the ability to inhibit a motor response as measured by SSRT improved with age, although this was of borderline significance (r = −0.28, P = 0.049). Not surprisingly, cortical thinning in the frontal, occipital, and parietal regions (Fig. 1) was correlated with better spatial and visual working memory performance in late childhood (Table 1). However, cortical thinning in these regions did not predict performance in the SST (Table 1).

Table 1

Relation of cortical thinning in late childhood with cognitive performance of the spatial working memory (SWM), delayed matching to sample (DMS), and stop-signal task (SST)

Cortical regions SWM DMS SSRT 
R-IFG 0.48*** −0.40** 0.27 
L-Cun/SPL 0.39** −0.39** 0.11 
L-IPL/LOC 0.35* −0.44*** 0.18 
R-IPS/LOC 0.33* −0.27 0.11 
L-vmPFC 0.33* −0.51** 0.19 
R-vmPFC 0.36* −0.38** 0.25 
Cortical regions SWM DMS SSRT 
R-IFG 0.48*** −0.40** 0.27 
L-Cun/SPL 0.39** −0.39** 0.11 
L-IPL/LOC 0.35* −0.44*** 0.18 
R-IPS/LOC 0.33* −0.27 0.11 
L-vmPFC 0.33* −0.51** 0.19 
R-vmPFC 0.36* −0.38** 0.25 

The table lists their Pearson's correlation coefficients.

SSRT, stop signal reaction time.

*P < 0.05; **P < 0.01; ***P < 0.001.

Functional Networks in Parallel with Cortical Thinning

Figure 2 shows the functional connectivity maps identified using rsfMRI seed-based correlation analysis considering the brain regions significantly associated with cortical thinning in late childhood as seeds (see Fig. 1). The R-IFG showed connections with bilateral medial superior frontal cortex (mSFC), orbital frontal cortex (OFC), IFG, insula, IPS/IPL, inferior temporal gyrus (ITG), fusiform, anterior cingulate cortex (ACC), right vmPFC, and precuneus (Fig. 2A). The left cuneus and superior parietal lobule (L-Cun/SPL) cortices were mainly connected with the sensorimotor and temporal regions of the cortex (Fig. 2B), including bilaterally superior, middle and inferior segments of central sulcus, paracentral gyrus, postcentral sulcus, lateral and medial occipital lobe, SPL, occipito-temporal junction (OTJ), posterior cingulate cortex (PCC), precuneus, superior temporal gyrus (STG), and left middle temporal gyrus (L-MTG). The left and right lateral occipital cortices (L-LOC/R-LOC) were connected bilaterally with lateral and medial occipital/parietal regions, except that R-LOC also showed connections with right parsopercularis (Fig. 2C and D). Similarly, bilateral vmPFC shared the same functional connections bilaterally with medial prefrontal, superior and orbital frontal cortices (mPFC/SFC/OFC), insula, supramarginal (SM), IPL, MTG, and PCC/precuneus (Fig. 2E and F).

Figure 2.

Functional connectivity maps associated with the seed region stated above each panel. On each panel, the top row shows the lateral view of the left and right cortical surfaces, while the bottom row shows the medial view of the left and right cortical surfaces. Vertexwise analyses were corrected for multiple comparisons (t-value = 2.68, P < 0.05 corrected).

Figure 2.

Functional connectivity maps associated with the seed region stated above each panel. On each panel, the top row shows the lateral view of the left and right cortical surfaces, while the bottom row shows the medial view of the left and right cortical surfaces. Vertexwise analyses were corrected for multiple comparisons (t-value = 2.68, P < 0.05 corrected).

The functional connectivity maps in Figure 2 parcellated the cortex into 15 regions per hemisphere (Fig. 3). Further graph modularity analysis identified 3 functional networks across these regions, including sensorimotor and auditory network (green in Fig. 4), executive-control network (blue in Fig. 4), and DMN (black in Fig. 4).

Figure 3.

Cortical parcellation based on the functional connectivity maps in Figure 2. Paracentral gyrus; middle central, middle segment of central sulcus; inferior central, inferior segment of central sulcus; STG, superior temporal gyrus; PCC, posterior cingulate cortex; mPFC, medial prefrontal cortex; SFC, superior frontal cortex; OFC, orbital frontal cortex; MTG, middle temporal gyrus; SPL, superior parietal lobule; IPS/IPL, intraparietal sulcus/inferior parietal lobule; IFG, inferior frontal gyrus; ITG, inferior temporal gyrus.

Figure 3.

Cortical parcellation based on the functional connectivity maps in Figure 2. Paracentral gyrus; middle central, middle segment of central sulcus; inferior central, inferior segment of central sulcus; STG, superior temporal gyrus; PCC, posterior cingulate cortex; mPFC, medial prefrontal cortex; SFC, superior frontal cortex; OFC, orbital frontal cortex; MTG, middle temporal gyrus; SPL, superior parietal lobule; IPS/IPL, intraparietal sulcus/inferior parietal lobule; IFG, inferior frontal gyrus; ITG, inferior temporal gyrus.

Figure 4.

Three functional networks were identified among the cortical regions in Figure 2, including sensorimotor and auditory network (green), executive-control network (blue), and default model network (black). The thickness of each edge denotes the connection strength between cortical regions. Paracentral gyrus; middle central, middle segment of central sulcus; inferior central, inferior segment of central sulcus; STG, superior temporal gyrus; PCC, posterior cingulate cortex; mPFC, medial prefrontal cortex; SFC, superior frontal cortex; OFC, orbital frontal cortex; MTG, middle temporal gyrus; SPL, superior parietal lobule; IPS/IPL, intraparietal sulcus/inferior parietal lobule; IFG, inferior frontal gyrus; ITG, inferior temporal gyrus.

Figure 4.

Three functional networks were identified among the cortical regions in Figure 2, including sensorimotor and auditory network (green), executive-control network (blue), and default model network (black). The thickness of each edge denotes the connection strength between cortical regions. Paracentral gyrus; middle central, middle segment of central sulcus; inferior central, inferior segment of central sulcus; STG, superior temporal gyrus; PCC, posterior cingulate cortex; mPFC, medial prefrontal cortex; SFC, superior frontal cortex; OFC, orbital frontal cortex; MTG, middle temporal gyrus; SPL, superior parietal lobule; IPS/IPL, intraparietal sulcus/inferior parietal lobule; IFG, inferior frontal gyrus; ITG, inferior temporal gyrus.

Efficiency of Functional Networks in Relations with Executive Functions and Age

To examine whether the aforementioned functional networks associated with age and cognitive performance in memory and/or response inhibition, we assessed relations (Pearson correlations) between cognitive measures and the graph properties of each functional network, including mean clustering coefficient (extent of local connectivity for a network) and characteristic path length (global efficiency of parallel information transfer). No significant associations of the graph properties of each network were found with age and cognitive measures (Table 2). Additionally, taking the 3 networks together, the mean clustering coefficient and characteristic path length were not related with age and the performance of working memory and response inhibition in children (Table 2). These findings suggest that the extent of local inter-connectivity and global efficiency of information transfer within and across sensorimotor and auditory, executive control, and DMN does not affect the performance on working memory and inhibition control in children.

Table 2

Graph properties of each functional network and all together in relation with age, spatial working memory (SWM), delayed matching to sample (DMS), and stop-signal task (SST)

Graph property Age SWM DMS SSRT 
Sensorimotor and auditory network 
 Mean clustering coefficient 0.18 0.04 0.03 −0.05 
 Characteristic path length 0.23 0.09 0.04 −0.12 
Executive control network 
 Mean clustering coefficient 0.04 −0.11 0.06 0.06 
 Characteristic path length −0.00 −0.11 0.06 0.08 
Default mode network 
 Mean clustering coefficient −0.22 0.02 −0.17 0.13 
 Characteristic path length −0.23 0.03 −0.12 0.01 
All 3 networks 
 Mean clustering coefficient 0.01 0.01 0.03 0.13 
 Characteristic path length 0.19 −0.03 0.07 −0.03 
Graph property Age SWM DMS SSRT 
Sensorimotor and auditory network 
 Mean clustering coefficient 0.18 0.04 0.03 −0.05 
 Characteristic path length 0.23 0.09 0.04 −0.12 
Executive control network 
 Mean clustering coefficient 0.04 −0.11 0.06 0.06 
 Characteristic path length −0.00 −0.11 0.06 0.08 
Default mode network 
 Mean clustering coefficient −0.22 0.02 −0.17 0.13 
 Characteristic path length −0.23 0.03 −0.12 0.01 
All 3 networks 
 Mean clustering coefficient 0.01 0.01 0.03 0.13 
 Characteristic path length 0.19 −0.03 0.07 −0.03 

The table lists their Pearson's correlation coefficients.

SSRT, stop signal reaction time.

*P < 0.05.

Regional Functional Connectivity in Relation with Executive Functions

Given the functional connectivity maps in Figure 2, we further examined whether connectivity strength (i.e., correlation of the rsfMRI signals in the seed region and the rest of the brain) was associated with working memory and response inhibition performance. Figure 5 illustrates that SWM and DMS were associated with common and distinct patterns of regionally specific connectivity strengths. First, the connections between the executive-control and sensorimotor networks and within the executive-control network predicted the cognitive performance of both SWM and DMS (Fig. 5A,D,G, Table 3). The lower connection strength between L-Cun/SPL and left middle central sulcus (L-MiddleCentral) predicted a smaller number of total errors in SWM and the higher total correct response score in DMS (Fig. 5A), while the higher connectivity strengths of L-IPL/LOC and R-IPS/LOC with left occipital temporal junction (L-OTJ) predicted better performance on both SWM and DMS tests (Fig. 5D,G). Secondly, distinct patterns of functional connections within the executive-control network were associated with the cognitive performance of SWM (Fig. 5F, Table 3) and DMS (Fig. 5D,E and H, Table 3), respectively. Stronger “dorsal” connections between L-IPL/LOC and right superior parietal lobule (R-SPL) predicted better performance on the SWM test (Fig. 5F), while stronger “ventral” connections between L-IPL/LOC and bilateral fusiform (Fig. 5D,E) and between R-IPS/LOC and right lingual gyrus (R-LG) (Fig. 5H) associated with better DMS performance. In addition, the connections between the executive-control network and DMN (Fig. 5B, Table 3) were associated with SWM performance, but the connections within DMN (Fig. 5I, Table 3) and between the executive-control and sensorimotor networks (Fig. 5C, Table 3) were related to DMS performance. The weaker connection between L-Cun/SPL and left precuneus (L-Precuneus) predicted better performance in the SWM task (Fig. 5B). The weaker connections between R-vmPFC and right superior frontal gyrus (R-SFG) within DMN (Fig. 5I) and between L-Cun/SPL and right middle central sulcus across the executive-control and sensorimotor networks (Fig. 5C) exhibited better DMS performance.

Table 3

MNI coordinates of peak functional connection-cognition relationships for memory performances

Seed Cluster Peak t-value Peak coordinates (MNI)
 
Total errors in spatial working memory 
 L-Cun/SPL L-MiddleCentral 4.25 −40 −16 55 
 L-IPL/LOC L-OTJ −4.65 −53 −59 
R-SPL −3.09 23 −64 63 
 R-IPS/LOC L-OTJ −4.11 −52 −56 −1 
 L-Cun/SPL L-Precuneus 4.27 −4 −43 63 
Total correct score in delayed matching to sample 
 L-Cun/SPL L-MiddleCentral −4.48 −36 −17 47 
R-MiddlePrecentral −3.78 42 −7 54 
R-MiddlePostcentral −3.69 42 −11 53 
 L-IPL/LOC L-OTJ 4.26 −47 −70  7 
L-Fusiform 4.53 −47 −55 −10 
R-Fusiform 3.72 43 −46 −5 
 R-IPS/LOC L-OTJ 3.66 −49 −71 
R-LG 3.42 11 −76 
 R-vmPFC R-SFG −4.20 15 25 57 
Seed Cluster Peak t-value Peak coordinates (MNI)
 
Total errors in spatial working memory 
 L-Cun/SPL L-MiddleCentral 4.25 −40 −16 55 
 L-IPL/LOC L-OTJ −4.65 −53 −59 
R-SPL −3.09 23 −64 63 
 R-IPS/LOC L-OTJ −4.11 −52 −56 −1 
 L-Cun/SPL L-Precuneus 4.27 −4 −43 63 
Total correct score in delayed matching to sample 
 L-Cun/SPL L-MiddleCentral −4.48 −36 −17 47 
R-MiddlePrecentral −3.78 42 −7 54 
R-MiddlePostcentral −3.69 42 −11 53 
 L-IPL/LOC L-OTJ 4.26 −47 −70  7 
L-Fusiform 4.53 −47 −55 −10 
R-Fusiform 3.72 43 −46 −5 
 R-IPS/LOC L-OTJ 3.66 −49 −71 
R-LG 3.42 11 −76 
 R-vmPFC R-SFG −4.20 15 25 57 

For locations of the clusters, please refer to Figure 5.

MiddleCentral, middle segment of central sulcus; OTJ, occipital temporal junction; SPL, superior parietal lobule; MiddlePrecentral, middle precentral gyrus; MiddlePostcentral, middle postcentral gyrus; LG, lingual gyrus; SFG, superior frontal gyrus.

Figure 5.

Relationships of the functional connectivity with spatial working memory (SWM) and delayed matching to sample tasks (DMS). (AC) Functional connectivity map of the left cuneus/superior parietal lobule (L-Cun/SPL) seed associated with total errors of SWM (cyan) and total correct of DMS (dark purple). (DF) and (GH), respectively, Functional connectivity map of the left inferior parietal lobule/lateral occipital cortex (L-IPL/LOC) and the right intraparietal sulcus/lateral occipital cortex (R-IPS/LOC) seeds associated with total errors of SWM (cyan) and total correct of DMS (dark purple). (I) Functional connectivity map of the right ventral medial prefrontal cortex (R-vmPFC) seed associated with total errors of SWM (cyan) and total correct of DMS (dark purple). The regions significantly associated with both SWM and DMS are in (yellow green). Positive and negative relationships are indicated by solid and dash circles, respectively. Vertexwise analyses were corrected for multiple comparisons (t-value = 2.68 for positive relationships and t-value = −2.68 for negative relationships; P < 0.05 corrected).

Figure 5.

Relationships of the functional connectivity with spatial working memory (SWM) and delayed matching to sample tasks (DMS). (AC) Functional connectivity map of the left cuneus/superior parietal lobule (L-Cun/SPL) seed associated with total errors of SWM (cyan) and total correct of DMS (dark purple). (DF) and (GH), respectively, Functional connectivity map of the left inferior parietal lobule/lateral occipital cortex (L-IPL/LOC) and the right intraparietal sulcus/lateral occipital cortex (R-IPS/LOC) seeds associated with total errors of SWM (cyan) and total correct of DMS (dark purple). (I) Functional connectivity map of the right ventral medial prefrontal cortex (R-vmPFC) seed associated with total errors of SWM (cyan) and total correct of DMS (dark purple). The regions significantly associated with both SWM and DMS are in (yellow green). Positive and negative relationships are indicated by solid and dash circles, respectively. Vertexwise analyses were corrected for multiple comparisons (t-value = 2.68 for positive relationships and t-value = −2.68 for negative relationships; P < 0.05 corrected).

Interestingly, our analysis revealed that the SSRT was associated with the connectivity strengths within the executive-control network (Fig. 6, Table 4). Better response inhibition (shorter SSRT) was predicted by the stronger connections between L-IPL/LOC and bilateral LG, as well as between R-IPS/LOC and R-LG, and the weaker connections between R-IFG and L-ITG, L-IPS, and R-Parsopercularis.

Table 4

MNI coordinates of peak functional connection-cognition relationships for stop signal tasks

Seed Cluster Peak t-value Peak coordinates (MNI)
 
Stop signal reaction time 
 L-IPL/LOC L-LG −3.61 −4 −54 
R-LG −3.18 −57 12 
 R-IPS/LOC R-LG −3.52 13 −39 −3 
 R-IFG L-ITG 3.18 −55 −70 
L-IPS 4.12 −23 −63 50 
R-Pasopercularis 3.22 35 34 
Seed Cluster Peak t-value Peak coordinates (MNI)
 
Stop signal reaction time 
 L-IPL/LOC L-LG −3.61 −4 −54 
R-LG −3.18 −57 12 
 R-IPS/LOC R-LG −3.52 13 −39 −3 
 R-IFG L-ITG 3.18 −55 −70 
L-IPS 4.12 −23 −63 50 
R-Pasopercularis 3.22 35 34 

For locations of the clusters, please refer to Figure 6.

LG, lingual gyrus; IFG, inferior frontal gyrus; ITG, inferior temporal gyrus; IPS, intraparietal sulcus.

Figure 6.

Relationships of the functional connectivity with the stop-signal reaction time (SSRT) for stop signal tasks (SST). (A and B) Functional connectivity map of the right inferior frontal gyrus (R-IFG) seed associated with SSRT (light purple). (C and D) and (E), respectively, Functional connectivity map of the left inferior parietal lobule/lateral occipital cortex (L-IPL/LOC) and the right intraparietal sulcus/lateral occipital cortex (R-IPS/LOC) seeds associated with SSRT. Positive and negative relationships are indicated by solid and dash circles, respectively. Vertexwise analyses were corrected for multiple comparisons (t-value = 2.68 for positive relationships and t-value = −2.68 for negative relationships; P < 0.05 corrected).

Figure 6.

Relationships of the functional connectivity with the stop-signal reaction time (SSRT) for stop signal tasks (SST). (A and B) Functional connectivity map of the right inferior frontal gyrus (R-IFG) seed associated with SSRT (light purple). (C and D) and (E), respectively, Functional connectivity map of the left inferior parietal lobule/lateral occipital cortex (L-IPL/LOC) and the right intraparietal sulcus/lateral occipital cortex (R-IPS/LOC) seeds associated with SSRT. Positive and negative relationships are indicated by solid and dash circles, respectively. Vertexwise analyses were corrected for multiple comparisons (t-value = 2.68 for positive relationships and t-value = −2.68 for negative relationships; P < 0.05 corrected).

Regional Functional Connectivity in Relation with Age and Cortical Thickness

Not surprisingly, the average strength of the regional functional connectivity in the clusters shown in Figures 5 and 6 was significantly correlated with the performance of SWM, DMS, and SST (Table 5). For most of the clusters in Figures 5 and 6, their connectivity strength was also associated with age. However, age was not associated with any of the connectivity strength in the clusters with the seed regions in prefrontal cortex (i.e., R-vmPFC and R-IFG, Table 5). Additionally, the connectivity strength in most of the clusters in Figures 5 and 6 was not related to cortical thickness averaged across the seed region of the corresponding cluster (Table 5).

Table 5

Pearson's correlation coefficients of regional functional connectivity with executive functions, age, and cortical thickness

Seed Cluster Executive functions Age Thickness 
Total errors in spatial working memory 
 L-Cun/SPL L-MiddleCentral 0.57*** −0.44** 0.30* 
 L-IPL/LOC L-OTJ −0.5*** 0.56*** −0.32* 
R-SPL −0.42** 0.24 −0.04 
 R-IPS/LOC L-OTJ −0.49*** 0.56*** −0.23 
 L-Cun/SPL L-Precuneus 0.47*** −0.35* 0.17 
Total correct score in delayed matching to sample 
 L-Cun/SPL L-MiddleCentral −0.57*** −0.36* 0.27 
R-MiddlePrecentral −0.5*** −0.26 0.07 
R-MiddlePostcentral −0.51*** −0.25 0.19 
 L-IPL/LOC L-OTJ 0.5*** 0.44** −0.36* 
L-Fusiform 0.49*** 0.29* −0.12 
R-Fusiform 0.5*** 0.21 −0.15 
 R-IPS/LOC L-OTJ 0.52*** 0.55*** −0.30* 
R-LG 0.44** 0.28* −0.14 
 R-vmPFC R-SFG −0.51*** −0.08 0.11 
Stop signal reaction time 
 L-IPL/LOC L-LG −0.44** 0.24 −0.01 
R-LG −0.42** 0.21 −0.004 
 R-IPS/LOC R-LG −0.45** 0.29* −0.07 
 R-IFG L-ITG 0.46*** −0.18 0.10 
L-IPS 0.51*** −0.28 0.06 
R-Parsopercularis 0.46*** 0.02 −0.04 
Seed Cluster Executive functions Age Thickness 
Total errors in spatial working memory 
 L-Cun/SPL L-MiddleCentral 0.57*** −0.44** 0.30* 
 L-IPL/LOC L-OTJ −0.5*** 0.56*** −0.32* 
R-SPL −0.42** 0.24 −0.04 
 R-IPS/LOC L-OTJ −0.49*** 0.56*** −0.23 
 L-Cun/SPL L-Precuneus 0.47*** −0.35* 0.17 
Total correct score in delayed matching to sample 
 L-Cun/SPL L-MiddleCentral −0.57*** −0.36* 0.27 
R-MiddlePrecentral −0.5*** −0.26 0.07 
R-MiddlePostcentral −0.51*** −0.25 0.19 
 L-IPL/LOC L-OTJ 0.5*** 0.44** −0.36* 
L-Fusiform 0.49*** 0.29* −0.12 
R-Fusiform 0.5*** 0.21 −0.15 
 R-IPS/LOC L-OTJ 0.52*** 0.55*** −0.30* 
R-LG 0.44** 0.28* −0.14 
 R-vmPFC R-SFG −0.51*** −0.08 0.11 
Stop signal reaction time 
 L-IPL/LOC L-LG −0.44** 0.24 −0.01 
R-LG −0.42** 0.21 −0.004 
 R-IPS/LOC R-LG −0.45** 0.29* −0.07 
 R-IFG L-ITG 0.46*** −0.18 0.10 
L-IPS 0.51*** −0.28 0.06 
R-Parsopercularis 0.46*** 0.02 −0.04 

The regional functional connectivity was computed as averaged connectivity strength in each cluster shown in Figure 5 for total errors in spatial working memory and total correct score in delayed matching to sample and in Figure 6 for stop signal reaction time. The thickness was computed as mean cortical thickness in the seed region.

See abbreviations in Tables 3 and 4.

*P < 0.05; **P < 0.01; ***P < 0.001.

Discussion

Our findings are consistent with previous studies showing that cortical thinning occurred bilaterally over the frontal, parietal, and occipital regions in late childhood (Sowell et al. 2004). We used these cortical regions as seeds and identified 3 functional networks, including the sensorimotor and auditory, executive control, and DMNs. The fine-tuned connections between specific cortical regions (i.e., functional connectivity strength) within or across the 3 functional networks, but not their topological organization (i.e., local and global efficiency of information transfer) predicted the performance of children on tests of working memory and response inhibition. The predictive ability of functional networks may add a unique contribution to cognitive development. Consistent with previous research (Hegarty et al. 2012), our findings suggest that the anatomical changes, reflected as cortical thinning, only partially predict the performance on tests of executive functions in late childhood.

Functional Connectivity Strength in Relation to Working Memory

Our study demonstrated that SWM and visual working memory shared common connections between the occipital and sensorimotor cortices in late childhood. Thus, fluctuations in resting BOLD signal of the occipital areas were highly correlated with those of the sensorimotor cortices, similar to the results reported by Nir et al. (2006) and Wang et al. (2008). Findings from task-fMRI studies show that the occipital areas and sensorimotor areas are jointly activated during mental imagery tasks (Mellet et al. 1996; Mazard et al. 2005). Therefore, Wang et al. (2008) postulated that mental imagery could be a candidate process underlying coherent spontaneous activities of the 2 cortical areas during the resting state. Our study further showed that such coherent fluctuations in resting BOLD signals of the occipital and sensorimotor cortices predict the performance of the SWM and DMS in children. Adequate suppression of activity between these 2 cortical areas was associated with better performance in the SWM and DMS tasks, suggesting that the ability to switch mental processes to visual and SWM processes is crucial for the performance of these cognitive tasks.

Our findings also showed that the dorsal and ventral visual pathways, respectively, were associated with performance on the SWM and DMS tasks in children. This finding lies in agreement with the organization of visual processing pathways reported in the monkey and human brains (Ungerleider et al. 1998). The locations of the occipitoparietal and occipitotemporal foci, respectively, are activated by spatial location matching and object matching. Nevertheless, our study did not reveal any connection of the prefrontal cortex, especially dorsolateral and ventrolateral prefrontal regions, with the occipital or parietal cortices in relation with the SWM and DMS performance in late childhood. This finding suggests that the performance of children on the SWM and DMS tasks may rely on the integration of externally driven perceptual information that demands bottom-up, but not on top-down processes to selectively focus upon relevant stimuli and ignore distracting stimuli.

Interestingly, our study further showed that the improvement of the SWM and DMS performance is not only associated with enhanced connections in areas of the dorsal and ventral visual streams, but also with attenuation of connections between occipitoparietal foci and DMN, and within DMN. This is in line with previous findings suggesting that better visual memory is associated with weaker connectivity between right ventromedial prefrontal cortex and superior frontal cortex within DMN (Hampson et al. 2010; Sambataro et al. 2010). Hence, adequate suppression of activity within the DMN is critical for allocation of the attentional resources necessary for the performance of a cognitive task.

Functional Connectivity Strength in Relation to Response Inhibition

Our study investigated resting-state functional connectivity in late childhood in relation to performance on the SST. We focused on the SSRT, which represents the rate of diffusion of a “stopping process.” Our study demonstrated that response inhibition in late childhood was associated with enhanced connectivity of the occipitopariental and a reduction in the connections between the right inferior prefrontal cortex and the intraparietal and inferior temporal cortices. This finding is consistent with numerous SST fMRI studies suggesting that the intraparietal sulcus and lateral occipital cortices are the primary cortical areas in the attentional processes, and that the right inferior prefrontal cortex is the cortical area responsible for the attentional processes involved in responding to an unexpected event (the “stop signal”) (Rubia et al. 2003; Sharp et al. 2010). Adult SST fMRI studies find greater activation in the superior frontal, precentral, and anterior cingulate cortices in the short compared with long SSRT group (Li et al. 2006). The effect was not apparent in our child study. One possible explanation for this discrepancy could be that the ability of children to inhibit a response may be triggered through automatic, bottom-up processing even though typically response inhibition is thought to be signaled by an external stop cue that provides a top-down signal to initiate the process. Existing fMRI studies support such an idea, suggesting that the right inferior frontal gyrus is involved in response inhibition regardless of whether its implementation occurred through top-down (triggered by the stop signal) or bottom-up (triggered by the association between a go stimulus and the stop signal) processes (Lenartowicz et al. 2011). Our finding emphasizes the role of the right inferior frontal gyrus in the bottom-up process in children. Further support for this idea comes from studies showing that the inferior frontal gyrus activity correlates with acquisition of associations between stimulus and response (Greicius et al. 2004) and the right inferior frontal cortex damage impairs learning regardless of inhibition (Wang et al. 2010). Hence, the connectivity of the right inferior frontal cortex with the intraparietal and inferior temporal cortices may comprise inhibition in late childhood through a bottom-up process that learns the association between a go stimulus and the stop signal.

Further Considerations

The bottom-up processes for performing working memory and response inhibition in late childhood may not come as a surprise in the view of the brain cellular maturation. Synaptogenesis and synapse elimination appear to be heterochronous in different cortical regions in humans (Huttenlocher and Dabholkar 1997). Synapse elimination in primary visual cortex appears to have started already in early childhood (Huttenlocher and Dabholkar 1997). In contrast, prefrontal cortex undergoes the largest synaptic overproduction and the slowest rate of synapse elimination in cerebral cortex and extends these processes to the third decade of life (Huttenlocher and Dabholkar 1997; Petanjek, Judas, Kostovic et al. 2008; Petanjek, Judas, Simic et al. 2011). This merges with the idea that emergence and elaboration of cortical function in the human appears to occur at different ages in different cortical regions. Functional development of prefrontal cortex, particularly more complex executive functions investigated in this study, is more gradual and at a later age than sensory cortex. Hence, we expect that the top-down processes for executive functions, as seen in adults (Li et al. 2006), may be gradually developed in the course of synapse overproduction and synaptic elimination in prefrontal cortex.

Existing studies suggested that cortical thinning during childhood is not only caused by a reduction in the size or number of neuron cell bodies and their synaptic processes but also an increase in the myelin coating of fibers in the lower cortical layers (Sowell, Peterson et al. 2003; Sowell, Thompson, Leonard et al. 2004). In prefrontal regions, synaptic overproduction reaches its maximum (Petanjek et al. 2008) and fractional anisotropy derived from diffusion tensor imaging increases as the reflection of the growth of myelination during late childhood (Barnea-Goraly et al. 2005). Hence, we speculate that during late childhood prefrontal cortical thinning would be mainly caused by increased proliferation of myelin, which may in turn strengthen prefrontal-related functional connections. However, our study revealed associations of prefrontal-related functional connectivity with executive functions but not age and its cortical thinning in late childhood. We suspect that functional connectivity might conversely shape anatomical connectivity through plastic changes driven by selective events that occur during development and evolution in childhood. This is also supported by evidence from a recent study suggesting that the topological organization of structural brain networks shifts toward random configuration in late childhood (Khundrakpam et al. 2012).

Limitations and Strengths

Unlike task-based fMRI, rsfMRI cannot reveal functional activations that respond to sequential external stimuli during cognitive tasks. Nevertheless, our study demonstrated that the functional organization in the resting brain associates with cognitive performance in children. This is particularly important in pediatric populations as rsfMRI requires a minimal cognitive burden on participants, and requires relatively little time in the scanner compared with task-fMRI studies. Our study demonstrated the feasibility using both structural MRI and rsfMRI for understanding complex patterns of the functional networks and thus yielding new insights into the brain organization underlying cognition and its plasticity in children. Hence, rsfMRI studies have the potential to reveal the dysfunctional network interactions that underlie the cognitive impairments associated with various clinical conditions in young children.

Conclusions

Our study demonstrated that cortical thinning only partially supports the development of executive functioning in late childhood. Fine-tuning but not network efficiency of the functional organization in parallel with the cortical development further explains cognitive performance in working memory and response inhibition in children. In particular, the executive functioning in children may much rely on the bottom-up processes, such as the integration of externally driven perceptual information that demands attention based on stimulus and learning the association between a go stimulus and the stop signal. The frontal connectivity related to top-down processes was found to be less involved in working memory and response inhibition in late childhood.

Funding

The work was supported by grant of Agency for Science, Technology and Research (A*STAR, SICS-09/1/1/001), the Young Investigator Award at the National University of Singapore (NUSYIA FY10 P07), the National University of Singapore MOE AcRF Tier 1, and Singapore Ministry of Education Academic Research Fund Tier 2 (MOE2012-T2-2-130).

Notes

Conflict of Interest: None declared.

References

Alderson
RM
Rapport
MD
Kofler
MJ
Attention-deficit/hyperactivity disorder and behavioral inhibition: a meta-analytic review of the stop-signal paradigm
J Abnormal Child Psychol
 , 
2007
, vol. 
35
 (pg. 
745
-
758
)
Amso
D
Casey
B
Beyond what develops when: neuroimaging may inform how cognition changes with development
Curr Dir Psychol Sci
 , 
2006
, vol. 
15
 (pg. 
24
-
29
)
Barnea-Goraly
N
Menon
V
Eckert
M
Tamm
L
Bammer
R
Karchemskiy
A
Dant
CC
Reiss
AL
White matter development during childhood and adolescence: A cross-sectional diffusion tensor imaging study
Cereb Cortex
 , 
2005
, vol. 
15
 (pg. 
1848
-
1854
)
Bray
S
Dunkin
B
Hong
DS
Reiss
AL
Reduced functional connectivity during working memory in Turner syndrome
Cereb Cortex
 , 
2011
, vol. 
21
 (pg. 
2471
-
2481
)
Cao
X
Cao
Q
Long
X
Sun
L
Sui
M
Zhu
C
Zuo
X
Zang
Y
Wang
Y
Abnormal resting-state functional connectivity patterns of the putamen in medication-naïve children with attention deficit hyperactivity disorder
Brain Res
 , 
2009
, vol. 
1303
 (pg. 
195
-
206
)
Chung
MK
Robbins
SM
Dalton
KM
Davidson
RJ
Alexander
AL
Evans
AC
Cortical thickness analysis in autism with heat kernel smoothing
Neuroimage
 , 
2005
, vol. 
25
 (pg. 
1256
-
1265
)
Chung
MK
Worsley
KJ
Nacewicz
BM
Dalton
KM
Davidson
RJ
General multivariate linear modeling of surface shapes using SurfStat
NeuroImage
 , 
2010
, vol. 
53
 (pg. 
491
-
505
)
Church
JA
Fair
DA
Dosenbach
NU
Cohen
AL
Miezin
FM
Petersen
SE
Schlaggar
BL
Control networks in paediatric Tourette syndrome show immature and anomalous patterns of functional connectivity
Brain
 , 
2009
, vol. 
132
 (pg. 
225
-
238
)
Dirani
M
Chan
YH
Gazzard
G
Hornbeak
DM
Leo
SW
Selvaraj
P
Zhou
B
Young
TL
Mitchell
P
Varma
R
, et al.  . 
Prevalence of refractive error in Singaporean Chinese children: The strabismus, amblyopia, and refractive error in young Singaporean Children (STARS) study
Invest Ophthalmol Vis Sci
 , 
2010
, vol. 
51
 (pg. 
1348
-
1355
)
Fair
DA
Cohen
AL
Dosenbach
NUF
Church
JA
Miezin
FM
Barch
DM
Raichle
ME
Petersen
SE
Schlaggar
BL
The maturing architecture of the brain's default network
Proc Natl Acad Sci USA
 , 
2008
, vol. 
105
 (pg. 
4028
-
4032
)
Fassbender
C
Murphy
K
Foxe
JJ
Wylie
GR
Javitt
DC
Robertson
IH
Garavan
H
A topography of executive functions and their interactions revealed by functional magnetic resonance imaging
Cogn Brain Res
 , 
2004
, vol. 
20
 (pg. 
132
-
143
)
Fischl
B
Dale
AM
Measuring the thickness of the human cerebral cortex from magnetic resonance images
Proc Natl Acad Sci USA
 , 
2000
, vol. 
97
 (pg. 
11050
-
11055
)
Fischl
B
Salat
DH
Busa
E
Albert
M
Dieterich
M
Haselgrove
C
van der Kouwe
A
Killiany
R
Kennedy
D
Klaveness
S
, et al.  . 
Whole brain segmentation: automated labeling of neuroanatomical structures in the human brain
Neuron
 , 
2002
, vol. 
33
 (pg. 
341
-
355
)
Fransson
P
SkiÖLd
B
EngstrÖM
M
Hallberg
B
Mosskin
M
ÅDen
U
Lagercrantz
H
Blennow
M
Spontaneous brain activity in the newborn brain during natural sleep—An fMRI study in infants born at full term
Pediatr Res
 , 
2009
, vol. 
66
 (pg. 
301
-
305
)
Gao
W
Zhu
H
Giovanello
KS
Smith
JK
Shen
D
Gilmore
JH
Lin
W
Evidence on the emergence of the brain's default network from 2-week-old to 2-year-old healthy pediatric subjects
Proc Natl Acad Sci USA
 , 
2009
, vol. 
106
 (pg. 
6790
-
6795
)
Greicius
MD
Srivastava
G
Reiss
AL
Menon
V
Default-mode network activity distinguishes Alzheimer's disease from healthy aging: evidence from functional MRI
Proc Natl Acad Sci USA
 , 
2004
, vol. 
101
 (pg. 
4637
-
4642
)
Hampson
M
Driesen
N
Roth
JK
Gore
JC
Constable
RT
Functional connectivity between task-positive and task-negative brain areas and its relation to working memory performance
Magn Reson Imaging
 , 
2010
, vol. 
28
 (pg. 
1051
-
1057
)
Hegarty
CE
Foland-Ross
LC
Narr
KL
Townsend
JD
Bookheimer
SY
Thompson
PM
Altshuler
LL
Anterior cingulate activation relates to local cortical thickness
NeuroReport
 , 
2012
, vol. 
23
 (pg. 
420
-
424
)
Hitch
GJ
Halliday
MS
Hulme
C
Voi
MEL
Routh
DA
Conway
A
Working memory in children [and discussion]
Philos Trans. R Soc Lond B Biol Sci
 , 
1983
, vol. 
302
 (pg. 
325
-
340
)
Humphries
MD
Gurney
K
Network ‘small-world-ness’: a quantitative method for determining canonical network equivalence
PLoS One
 , 
2008
, vol. 
3
 pg. 
e0002051
 
Huttenlocher
PR
Dabholkar
AS
Regional differences in synaptogenesis in human cerebral cortex
J Comp Neurol
 , 
1997
, vol. 
387
 (pg. 
167
-
178
)
Jurado
MB
Rosselli
M
The elusive nature of executive functions: A review of our current understanding
Neuropsychol Rev
 , 
2007
, vol. 
17
 (pg. 
213
-
233
)
Khundrakpam
BS
Reid
A
Brauer
J
Carbonell
F
Lewis
J
Ameis
S
Karama
S
Lee
J
Chen
Z
Das
S
, et al.  . 
Developmental changes in organization of structural brain networks
Cereb Cortex
 , 
2012
Latora
V
Marchiori
M
Efficient behavior of small-world networks
Phys Rev Lett
 , 
2001
, vol. 
87
 pg. 
198701
 
Lenartowicz
A
Verbruggen
F
Logan
GD
Poldrack
RA
Inhibition-related activation in the right inferior frontal gyrus in the absence of inhibitory cues
J Cogn Neurosci
 , 
2011
, vol. 
23
 (pg. 
3388
-
3399
)
Lenroot
RK
Giedd
JN
Brain development in children and adolescents: Insights from anatomical magnetic resonance imaging
Neurosci Biobehav Rev
 , 
2006
, vol. 
30
 (pg. 
718
-
729
)
Li
CS
Huang
C
Constable
RT
Sinha
R
Imaging response inhibition in a stop-signal task: Neural correlates independent of signal monitoring and post-response processing
J Neurosci
 , 
2006
, vol. 
26
 (pg. 
186
-
192
)
Luciana
M
Nelson
CA
Assessment of neuropsychological function through use of the Cambridge Neuropsychological Testing Automated Battery: performance in 4- to 12-year-old children
Dev Neuropsychol
 , 
2002
, vol. 
22
 (pg. 
595
-
624
)
Mazard
A
Laou
L
Joliot
M
Mellet
E
Neural impact of the semantic content of visual mental images and visual percepts
Cogn Brain Res
 , 
2005
, vol. 
24
 (pg. 
423
-
435
)
Mellet
E
Tzourio
N
Crivello
F
Joliot
M
Denis
M
Mazoyer
B
Functional anatomy of spatial mental imagery generated from verbal instructions
J Neurosci
 , 
1996
, vol. 
16
 (pg. 
6504
-
6512
)
Miller
EK
Erickson
CA
Desimone
R
Neural mechanisms of visual working memory in prefrontal cortex of the macaque
J Neurosci
 , 
1996
, vol. 
16
 (pg. 
5154
-
5167
)
Muftuler
LT
Davis
EP
Buss
C
Head
K
Hasso
AN
Sandman
CA
Cortical and subcortical changes in typically developing preadolescent children
Brain Res
 , 
2011
, vol. 
1399
 (pg. 
15
-
24
)
Nir
Y
Hasson
U
Levy
I
Yeshurun
Y
Malach
R
Widespread functional connectivity and fMRI fluctuations in human visual cortex in the absence of visual stimulation
NeuroImage
 , 
2006
, vol. 
30
 (pg. 
1313
-
1324
)
Petanjek
Z
Judas
M
Kostovic
I
Uylings
HB
Lifespan alterations of basal dendritic trees of pyramidal neurons in the human prefrontal cortex: A layer-specific pattern
Cereb Cortex
 , 
2008
, vol. 
18
 (pg. 
915
-
929
)
Petanjek
Z
Judas
M
Simic
G
Rasin
MR
Uylings
HB
Rakic
P
Kostovic
I
Extraordinary neoteny of synaptic spines in the human prefrontal cortex
Proc Natl Acad Sci USA
 , 
2011
, vol. 
108
 (pg. 
13281
-
13286
)
Pickering
SJ
The development of visuo-spatial working memory
Memory
 , 
2001
, vol. 
9
 (pg. 
423
-
432
)
Qiu
A
Rosenau
BJ
Greenberg
AS
Hurdal
MK
Barta
P
Yantis
S
Miller
MI
Estimating linear cortical magnification in human primary visual cortex via dynamic programming
Neuroimage
 , 
2006
, vol. 
31
 (pg. 
125
-
138
)
Rakic
P
Bourgeois
JP
Goldman-Rakic
PS
Synaptic development of the cerebral cortex: implications for learning, memory, and mental illness
Progress Brain Res
 , 
1994
, vol. 
102
 (pg. 
227
-
243
)
Rubia
K
Smith
AB
Brammer
MJ
Taylor
E
Right inferior prefrontal cortex mediates response inhibition while mesial prefrontal cortex is responsible for error detection
Neuroimage
 , 
2003
, vol. 
20
 (pg. 
351
-
358
)
Sambataro
F
Murty
VP
Callicott
JH
Tan
H-Y
Das
S
Weinberger
DR
Mattay
VS
Age-related alterations in default mode network: Impact on working memory performance
Neurobiol Aging
 , 
2010
, vol. 
31
 (pg. 
839
-
852
)
Seeley
WW
Menon
V
Schatzberg
AF
Keller
J
Glover
GH
Kenna
H
Reiss
AL
Greicius
MD
Dissociable intrinsic connectivity networks for salience processing and executive control
J Neurosci
 , 
2007
, vol. 
27
 (pg. 
2349
-
2356
)
Sharp
DJ
Bonnelle
V
De Boissezon
X
Beckmann
CF
James
SG
Patel
MC
Mehta
MA
Distinct frontal systems for response inhibition, attentional capture, and error processing
Proc Natl Acad Sci USA
 , 
2010
, vol. 
107
 (pg. 
6106
-
6111
)
Song
M
Zhou
Y
Li
J
Liu
Y
Tian
L
Yu
C
Jiang
T
Brain spontaneous functional connectivity and intelligence
NeuroImage
 , 
2008
, vol. 
41
 (pg. 
1168
-
1176
)
Sowell
ER
Peterson
BS
Thompson
PM
Welcome
SE
Henkenius
AL
Toga
AW
Mapping cortical change across the human life span
Nat Neurosci
 , 
2003
, vol. 
6
 (pg. 
309
-
315
)
Sowell
ER
Thompson
PM
Holmes
CJ
Jernigan
TL
Toga
AW
In vivo evidence for post-adolescent brain maturation in frontal and striatal regions
Nat Neurosci
 , 
1999
, vol. 
2
 (pg. 
859
-
861
)
Sowell
ER
Thompson
PM
Leonard
CM
Welcome
SE
Kan
E
Toga
AW
Longitudinal mapping of cortical thickness and brain growth in normal children
J Neurosci
 , 
2004
, vol. 
24
 (pg. 
8223
-
8231
)
Sowell
ER
Trauner
DA
Gamst
A
Jernigan
TL
Development of cortical and subcortical brain structures in childhood and adolescence: a structural MRI study
Dev Med Child Neurol
 , 
2002
, vol. 
44
 (pg. 
4
-
16
)
Spear
LP
The adolescent brain and age-related behavioral manifestations
Neurosci Biobehav Rev
 , 
2000
, vol. 
24
 (pg. 
417
-
463
)
Swartz
BE
Simpkins
F
Halgren
E
Mandelkern
M
Brown
C
Krisdakumtorn
T
Gee
M
Visual working memory in primary generalized epilepsy: an 18FDG-PET study
Neurology
 , 
1996
, vol. 
47
 (pg. 
1203
-
1212
)
Tillman
CM
Thorell
LB
Brocki
KC
Bohlin
G
Motor response inhibition and execution in the stop-signal task: Development and relation to ADHD behaviors
Child Neuropsychol
 , 
2008
, vol. 
14
 (pg. 
42
-
59
)
Uddin
LQ
Supekar
K
Menon
V
Typical and atypical development of functional human brain networks: insights from resting-state fMRI
Front Syst Neurosci
 , 
2010
, vol. 
4
 pg. 
21
 
Ungerleider
LG
Courtney
SM
Haxby
JV
A neural system for human visual working memory
Proc Natl Acad Sci USA
 , 
1998
, vol. 
95
 (pg. 
883
-
890
)
Urben
S
Van der Linden
M
Barisnikov
K
Development of the ability to inhibit a prepotent response: Influence of working memory and processing speed
Br J Dev Psychol
 , 
2011
, vol. 
29
 (pg. 
981
-
998
)
van den Heuvel
MP
Stam
CJ
Kahn
RS
Hulshoff Pol
HE
Efficiency of functional brain networks and intellectual performance
J Neurosci
 , 
2009
, vol. 
29
 (pg. 
7619
-
7624
)
Vogel
A
Power
J
Petersen
S
Schlaggar
B
Development of the brain's functional network architecture
Neuropsychol Rev
 , 
2010
, vol. 
20
 (pg. 
362
-
375
)
Vuontela
V
Steenari
MR
Carlson
S
Koivisto
J
Fjallberg
M
Aronen
ET
Audiospatial and visuospatial working memory in 6–13 year old school children
Learn Mem
 , 
2003
, vol. 
10
 (pg. 
74
-
81
)
Wang
L
Negreira
A
LaViolette
P
Bakkour
A
Sperling
RA
Dickerson
BC
Intrinsic interhemispheric hippocampal functional connectivity predicts individual differences in memory performance ability
Hippocampus
 , 
2010
, vol. 
20
 (pg. 
345
-
351
)
Wang
K
Jiang
T
Yu
C
Tian
L
Li
J
Liu
Y
Zhou
Y
Xu
L
Song
M
LI
K
Spontaneous activity associated with primary visual cortex: a resting-state FMRI study
Cerebral Cortex
 , 
2008
, vol. 
18
 (pg. 
697
-
704
)
Zar
JH
Biostatistical Analysis
 , 
1996
3rd ed. Upper Saddle River, NJ: Prentice Hall
Zhong
J
Qiu
A
Multi-manifold diffeomorphic metric mapping for aligning cortical hemispheric surfaces
NeuroImage
 , 
2010
, vol. 
49
 (pg. 
355
-
365
)