Relationship Between White Matter Abnormalities and Neuropsychological Measures in Children With ADHD

Abstract

Objective: Using Diffusion Tensor Imaging (DTI), to investigate microstructural white matter differences between ADHD and typically developing children (TDC), and their association with inhibition and working memory performance usually impaired in ADHD. Method: Fractional anisotropy (FA) and mean diffusivity (MD) were estimated in 36 noncomorbid children with a Diagnostic and Statistical Manual of Mental Disorders (4th ed., text rev.; DSM-IV-TR) diagnosis of combined type ADHD and 20 TDC. Correlations between FA/MD and Stop Signal Task and N-Back performance parameters were computed. Results: Working memory performance was significantly associated with MD in the superior longitudinal fasciculus (SLF) and the cingulum in the ADHD group. No between-group differences in FA/MD reached significance, after controlling for between-group head motion differences. Conclusion: The association between white matter integrity in the cingulum and the SLF and working memory performance confirms previous studies. Our results also show that when critical conditions are controlled (age, comorbidity, head motion), no ADHD-related structural abnormality (FA/MD) are observed, in line with prior suggestions.

Keywords

ADHD inhibition working memory DTI TBSS

Attention deficit hyperactivity disorder (ADHD) is a common neurodevelopmental disorder (estimated worldwide-pooled prevalence > 5%), characterized by impairing levels of inattention, impulsivity, and/or hyperactivity, which may persist into adulthood (American Psychiatric Association [APA], 2000, 2013; Thomas, Sanders, Doust, Beller, & Glasziou, 2015). At the neuropsychological level, populations with ADHD perform worse than typically developing peers on a variety of neurocognitive domains, including working memory and response inhibition (Pievsky & McGrath, 2017). One of the working memory tasks in which children with ADHD exhibit decreased performance when compared with typically developing children (TDC) is the N-Back task (Bechtel et al., 2012; Shallice et al., 2002). In this task, participants have to monitor a series of stimuli and respond whenever a presented stimulus is the same as the one presented at a preceding trial, for example, N-2. Likewise, motor response inhibition was often investigated in ADHD using the Stop Signal Task (SST), in which already initiated responses to Go stimuli must be inhibited upon the delayed presentation of a Stop stimulus (Logan, 1994; Logan, Van Zandt, Verbruggen, & Wagenmakers, 2014). One meta-analysis reported medium effect sizes (g = .62) for SST comparisons between children with ADHD and TDC, suggesting that approximately two thirds of children with ADHD will exhibit a performance below the mean of TDC groups (Lipszyc & Schachar, 2010). Longer stop-signal reaction times (SSRT) in patients with ADHD have been interpreted as reflecting reduced inhibition speed in models emphasizing the role of response inhibition deficits in ADHD (Lipszyc & Schachar, 2010; Pievsky & McGrath, 2017).

Besides behavioral paradigms, neuroimaging tools offer useful complementary information to disentangle competing models of ADHD. In particular, structural brain imaging may contribute to identify whether core brain regions involved in response inhibition versus working memory are altered in patients with ADHD. Diffusion Tensor Imaging (DTI) is a noninvasive in vivo technique that characterizes the microstructural properties of brain white matter tracts (Basser, 1995). Diffusion features different scalar measures, corresponding to different geometrical properties of the diffusion tensor. The most commonly used DTI-derived index is fractional anisotropy (FA), defined as the fraction of the magnitude tensor that can be attributed to anisotropic diffusion, ranging from 0 (maximally isotropic) to 1 (maximally anisotropic; Pierpaoli & Basser, 1996). Low FA was associated with impaired myelination, axonal density, and fiber organization and integrity (Alba-Ferrara & de Erausquin, 2013). Another commonly used DTI parameter is mean diffusivity (MD), which reflects the overall diffusivity in a particular voxel. High MD values potentially indicate tissue damage and lack of directionality. One analytic approach to extract and compare diffusion parameters between groups is Tract-Based Spatial Statistics (TBSS), which is based on the construction of a mean FA tract skeleton in which all participants’ FA data are projected (Smith et al., 2006). By doing so, TBSS overcomes some drawbacks of other approaches such as voxel-based analysis (VBA), by limiting the risk for registration errors and removing the need for smoothing.

Over the last few years, the number of DTI studies in ADHD has progressively increased and three major meta-analyses synthesized their results (Aoki, Cortese, & Castellanos, 2017; Chen et al., 2016; van Ewijk, Heslenfeld, Zwiers, Buitelaar, & Oosterlaan, 2012). The most recent meta-analysis by Aoki et al. (2017) was performed separately on 14 VBA and 13 TBSS studies. The VBA meta-analysis evidenced both increased and decreased FA in participants with ADHD in six clusters including the cingulum, the corpus callosum (CC) and the left inferior-fronto occipital fasciculus (IFOF), whereas TBSS studies evidenced lower FA in ADHD in five clusters, that is, the isthmus and the posterior midbody of the CC, the right IFOF, the left inferior longitudinal fasciculus (ILF) and the right superior longitudinal fasciculus (SLF). Overall, multiple white matter tracks have been linked to ADHD with the most common finding being an atypical interhemispheric CC connection.

At the methodological level, it should be noticed that most previous studies did not control for potential between-group differences in head motion, thus not accounting for the known increased head motion in ADHD samples (Aoki et al., 2017). Between-group differences in head motion may result in spurious brain imaging findings, especially in diffusion imaging studies (Yendiki, Koldewyn, Kakunoori, Kanwisher, & Fischl, 2014). In particular, Aoki et al. (2017) noticed that out of six recent TBSS studies in which head motion did not significantly differ between groups, four studies did not disclose between-group differences in FA. Thus, future studies should control for this confounding factor. It should also be noted that an effect of age was evidenced in the meta-analysis by Chen et al. (2016), who found that decreased FA in the splenium of the CC was negatively associated with the mean age of patients with ADHD.

In the recent years, authors have investigated the relationship between DTI-derived indices (generalized fractional anisotropy [GFA], i.e., an analogous of FA in Diffusion Spectrum Imaging [DSI]) and neuropsychological task parameters assessing different cognitive functions, including inhibition and working memory. Regarding impulsivity and inhibition, some studies evidenced an association between task performance and DTI parameters in brain regions including the CC, the anterior, posterior and superior corona radiata, the SLF, the sagittal stratum, and the cingulum (Chiang, Chen, Lo, Tsen, & Gau, 2015; Onnink et al., 2015; Rossi et al., 2015; Wu et al., 2016). To the best of our knowledge, no ADHD studies have examined the relationship between DTI-derived indices and the SSRT, one of the key outcome measures of motor response inhibition in the SST. In healthy adults and children, correlations between SSRT and DTI-derived indices (FA/RD) were evidenced in the white matter underlying the right inferior frontal gyrus (IFG), the right pre-supplementary motor area (preSMA), and the subthalamic nucleus (STN; Coxon, Van Impe, Wenderoth, & Swinnen, 2012; Madsen et al., 2010): as well as in the fibers connecting the preSMA/SMA with the STN and the striatum, and in a tract connecting the IFG with the STN (Coxon et al., 2012; King et al., 2012). King et al. (2012) also found that tracts whose integrity was associated with SSRT passed through the internal capsule, the anterior corona radiata, the superior corona radiata, the basal ganglia, or the spinal cord. They also reported an association between SSRT and FA/RD in the CC, the inferior fronto-occipital tract, the precentral gyrus, and the optic radiation. Regarding working memory, although some authors failed to evidence any association with DTI/DSI derived indices (Li et al., 2010; Onnink et al., 2015; Wu et al., 2016), other studies in children and adolescents with ADHD reported a correlation with GFA in different tracts such as the left orbitofrontal fiber tract (Shang, Wu, Gau, & Tseng, 2012), the bilateral SLF, and the left arcuate fasciculus (Chiang, Chen, Shang, Tseng, & Gau, 2016). To the best of our knowledge, no study investigated the association between working memory as measured by the N-Back task (Owen, McMillan, Laird, & Bullmore, 2005) and white matter regional integrity in ADHD. In healthy children and adults, the microstructural integrity of multiple tracts including the cingulum, the CC, and the SLF have been related to working memory performance in various tasks (Nagy, Westerberg, & Klingberg, 2004; Peters et al., 2012; Schulze et al., 2011; Short et al., 2013; Takahashi et al., 2010; Vestergaard et al., 2011). In particular, the fiber tracts connecting prefrontal and parietal regions such as the SLF seem to play an important role in working memory.

To sum up, a growing DTI literature evidenced white matter microstructure differences between patients with ADHD and healthy controls, especially in the CC. However, it was deemed too premature to draw conclusions about structural connectivity in ADHD (Aoki et al., 2017). Indeed, DTI studies examining differences between ADHD and healthy participants widely vary in their sample size, age range, medication status, clinical subtype, comorbidities, head motion correction, and data analysis methodology (whole-brain VBA, TBSS, region-of-interest approaches). In particular, the inclusion of patients with comorbid psychiatric or neurodevelopmental disorders may hinder the characterization of white matter anomalies that would be specific to pure ADHD cases. In addition, only a limited number of studies have investigated the relationship between DTI-derived indices and working memory or response inhibition, two functions typically impaired in ADHD.

In the present DTI study, we studied white matter microstructure in children with a Diagnostic and Statistical Manual of Mental Disorders (4th ed., text rev.; DSM-IV-TR; American Psychiatric Association [APA], 2000) diagnosis of combined type ADHD without comorbidities and in TDC using the TBSS approach. We also aimed at probing the relationships between white matter track changes and inhibition and working memory performance. To do so, we analyzed between-group differences in FA and MD parameters, and the association between these DTI measures and behavioral performance levels in the N-Back and the SST. Analyses were performed both at the whole-brain level and in four preselected fiber tracts: the CC, the cingulum, the sagittal stratum, and the SLF.

Method

Participants

In total, 58 children (32 males), aged 8 to 12.5 years, participated in this study approved by the Université Libre de Bruxelles (ULB) Ethics Committee (reference: P2007/332/B40620072950) and performed in accordance with the ethical standards of the 1964 Declaration of Helsinki. Each child and her or his parents gave written consent to participate. Children with ADHD were recruited from the neuropediatric outpatient clinic in Erasme Hospital, ULB, Belgium. TDC were recruited from local schools in Brussels or personal networks. Two participants with ADHD were excluded after visual inspection of DTI data quality due to movement artifacts in nondiffusion weighted images. Final analyses were conducted on 56 children (31 males; mean age M = 122.98 months; standard deviation [SD] = 15.41 months): 36 children fulfilling the DSM-IV-TR criteria for ADHD combined type (APA, 2000) and 20 TD children. Intelligence quotient (IQ) estimates were obtained with the age-appropriate Wechsler Abbreviated Scale of Intelligence (WASI; Wechsler, 1999). All participants had an IQ higher than 80. The ADHD group had a mean IQ of 105.53 (SD = 10.85), and the TD group had a mean IQ of 109.39 (SD = 9.22). Diagnosis for ADHD was based on clinical features including typical history and behavioral report. The Kiddie Schedule for Affective Disorders and Schizophrenia for School Aged Children-Present and Lifetime Version (K-SADS-PL; Endicott & Spitzer, 1978) was completed at screening for each participant to establish the diagnosis according to DSM-IV-TR criteria in children with ADHD and to ensure that TDC presented no psychiatric condition. ADHD severity in children with ADHD was measured with the ADHD Rating Scale–IV (RS-IV; DuPaul, 1998). Higher scores indicate more severe impairment (ADHD group: M = 38.18, SD = 6.14, 3 cases missing). In total, 16 participants with ADHD (44.44%) were under medication, receiving between 0.3 and 0.5 mg of methylphenidate per kilo (per dose, 3 times a day), for a maximal duration of 3 years. All other participants did not receive any medication. Exclusion criteria for all children were presence of a psychiatric condition other than ADHD, history of prematurity, current or past medical or neurological disorder, contraindication to magnetic resonance imaging (MRI), and IQ estimate below 80. Poor task performance was considered as an exclusion criterion (see below).

Neuropsychological Measures

Motor response inhibition: SST

As mentioned above, the SST measures the ability to suppress a previously triggered motor response to a go signal, when it is unpredictably followed by a stop signal shortly after (see Figure 1). The SST was performed during a functional MRI acquisition (fMRI data are published elsewhere and not discussed here, Massat et al., 2016). Go signals were successive arrows pointing either to the left (n = 100) or to the right (n = 100). Mean inter-stimulus interval ranged from 2,500 to 3,750 ms. Participants were instructed to respond as accurately and quickly as possible by pressing the right or left button response with their right or left thumb according to the direction of the arrow. In 20% of the trials, a stop signal was presented shortly after the go signal (20 after a right-oriented go signal, 20 after a left-oriented go signal). The stop signal was an arrow pointing upward. Stop trials were pseudo-randomly interspersed with Go trials. Initially, the stop trial was displayed after a 250 ms Stop Signal Delay (SSD), that is, 250 ms after the go signal. Subsequent SSDs were then adapted according to the participant’s performance using a horse race model (Logan, 1994; Logan et al., 2014). If inhibition was successful, then the SSD was made longer by 50 ms to make inhibition of the ongoing response more difficult. If inhibition failed, then the SSD was made shorter by 50 ms to facilitate inhibition. Participants were instructed to attempt canceling their ongoing motor response (i.e., not responding) when the stop signal appeared. They were also told not to wait for the stop stimulus to appear, and that it was perfectly normal not being able to cancel their responses on all stop trials. Task parameters in the SST are the percentage of correctly inhibited stop trials, the number of omissions on Go trials, the SSD, the mean reaction time (MRT) on go signals, and the SSRT. SSRT, which is here the main outcome measure, is the time needed by the participant to process the stop signal and successfully withhold the already initiated response (Logan, 1994; Logan et al., 2014). SSRT is computed by subtracting the mean SSD (i.e., the mean SSD at which the participant successfully inhibited 50% of the stop trials) from the MRT. SSRT provides an individual index of inhibitory control (Logan, 1994; Logan et al., 2014), with longer SSRTs indicating poorer response inhibition. SST performance scores were available for 55 participants (20 TDC and 35 ADHD). Based on Congdon et al. (2012), a participant meeting any of the following criteria for the SST was excluded: percentage of inhibition on Stops trials inferior to 25% or superior to 75%, percentage of correct responses on Go trials inferior to 60%, percentage of incorrect direction on Go trials superior to 10%, and an SSRT estimate negative or inferior to 50 ms. In total, 16 participants were excluded because they did not meet this criteria and data were analyzed in the 39 remaining participants (17 TDC and 22 ADHD). Comparisons between groups were computed using Welch’s t test (or Mann–Whitney-Wilcoxon Test if the distribution was not normal).

Figure 1.

Schematic illustration of the SST.

Working memory updating: N-Back task

The N-Back task was also performed during an fMRI acquisition (fMRI data not discussed here, see Massat et al., 2012). All participants were trained on the whole task outside of the fMRI environment once before scanning; results from the task performed during the fMRI acquisition were used here. The N-Back task features two different conditions (see Figure 2). In both conditions, stimuli were black numbers (Arial font, size 74) displayed on a white background on the center of the screen, successively presented in pseudo-random order. In the detection/control 0-Back (N0) condition, participants had to press a button with the right hand whenever the number “2” was displayed. In the 2-Back (N2) condition, participants had to press the button when the displayed number was identical to the number displayed two trials before. Number of stimuli and motor responses were identical in the two conditions, but active manipulation of information was requested in the N2 back condition only. Participants were administered five blocks in the N0 condition, alternated with five blocks in the N2 condition. Each block consisted of a sequence of 30 trials (including 10 targets for which a response had to be provided), each displayed for 1,750 ms with an inter-stimulus interval of 250 ms. Each block was followed by a resting period of random duration ranging 11 to 16 s, during which the instruction relative to the forthcoming condition was displayed (i.e., either “number 2” [N0] or “same than two numbers before” [N2]). A fixation cross replaced the instruction 2.5 s before the start of a novel series of 30 numbers. Corrected accuracy scores (hits – false detections/2) were obtained in the N2 and N0 conditions. The main outcome measure was the Updating Process (UP = N0-N2 corrected scores). This measure reflects the participant’s ability to discriminate target from nontargets, removing the potential effect of attentional performance. N-Back performance scores were available for 53 participants (18 TDC and 35 ADHD). Comparisons between groups were computed using Welch’s t test (or Mann–Whitney-Wilcoxon Test if the distribution was not normal). Because we examined three different measures, a Bonferroni correction for multiple comparisons was further applied based on the number of variables measures for each task. The statistical threshold was set at .05/5 = .0001 for the Stop task, and .05/3 = .017 for the N-Back task.

Figure 2.

Schematic illustration of the N-Back.

Diffusion Imaging

MRI data acquisition

DTI acquisitions consisted of two transverse single shot echo planar imaging DTI scans (60 diffusion-weighted volumes with different noncollinear diffusion encoding directions, with b-factor 1,000 s/mm² and 10 diffusion-unweighted volumes with b-factor 0 s/mm²; parallel imaging SENSE factor 2; flip angle 90°; 65 slices of 2 mm; no gap; 100 x 100 acquisition matrix; reconstruction matrix 128 x 128; field of view (FOV) 200 mm; echo time (TE) 76 ms; repetition time (TR) 6,750 ms).

MRI preprocessing

Functional Magnetic Resonance Imaging of the Brain (FMRIB)’s Diffusion Toolbox (FDT; http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/FDT) was used for eddy current correction and elimination of motion artifacts. Tensor-based registration was performed using the DTI-TK (http://dti-tk.sourceforge.net) software package, which uses the full diffusion tensor information to drive the registration and improve the alignment of white matter structures. DTI-TK was shown to align white matter pathways better than the scalar-based registration methods used in the standard TBSS pipeline, which has limitations (Keihaninejad et al., 2012). FA and MD data from each participant were further processed and analyzed using the TBSS tool available in the FSL toolbox (Smith et al., 2006). A skeleton of white matter was generated by thresholding the mean FA map (FA N 0.2) and representing a single line running down the centers of all the common white matter fibers (Smith et al., 2006). Each participant’s normalized FA and MD images were then projected onto this common skeleton to minimize any residual misalignment of tracts.

TBSS

To account for the effects of potential group differences in head motion during scanning, we computed a head motion index for each participant, that is, the mean absolute inter-volume displacement with respect to the first volume acquired. This head motion index was used as a confound covariate in all analyses to ascertain that group differences were not merely due to head motion. The JHU-ICBM-tract atlas (Wakana et al., 2007) was used to define regions of interest (ROIs) for major fiber bundles that were previously shown important in relationship with ADHD, working memory, and/or inhibition. The atlas was linearly and nonlinearly registered to the study final FA template using FLIRT and FNIRT from FSL. ROIs, which included the CC (genu, body, and splenium), bilateral superior longitudinal fascicles, sagittal stratum (encompassing the inferior longitudinal fascicles), and cingulum bundles were extracted from the template using FSLmaths. Statistical analyses included a between-group comparison and correlations between inhibition performance, working memory performance, and DTI metrics in the whole sample and separately in the ADHD and the TDC groups. For correlation analysis, the SST score used for inhibition performance was the SSRT, and the N-Back score used for working memory performance was the UP. We also performed a correlation between DTI metrics and the accuracy score at the N0-Back, which corresponds to the vigilance/control condition. For each score, we excluded participants with an outlying performance using the interquartile range. Statistics were computed based on a nonparametric voxel-wise analysis of the skeleton-voxels, in the whole skeletonized template as well as within our four ROIs (Smith & Nichols, 2009). The model was estimated using the “randomize” tool available in FSL, adjusting for age, gender, and head motion. Because IQ is typically associated with ADHD and does not meet the requirements for a covariate (Dennis et al., 2009; Miller & Chapman, 2001), all analyses were conducted without IQ as a covariate. The number of permutations was set to 5,000 (Nichols & Holmes, 2002), and results were corrected for multiple comparisons using the threshold-free cluster enhancement (TFCE) method, and results were considered significant (p-FWE-corrected < .05). Because we examined three different measures, a Bonferroni correction for multiple comparison was further applied to the results of the correlational analyses (significance set at p-FWE-corrected < .05/3 = .017).

Results

Demographic Data and Head Motion

Demographic data and head motion for ADHD and TDC are reported in Table 1. Welch’s t tests disclosed a trend for older age in children with ADHD compared with TDC (p = .06). Gender was not significantly related to diagnosis (p = .55). IQ was available for 33 ADHD and 18 TDC. All participants had an IQ > 80, and no significant difference emerged between groups. Finally, there were no significant differences between groups regarding head motion.

Table 1.

Demographic Data and Head Motion of Analyzed Participants.

	ADHD (n = 36)		TDC (n = 20)		χ²	df	p
	M/F		M/F		χ²	df	p
Gender	21/15		10/10		0.36	1	.55
	M	SD	M	SD	T	df	p
Age (months)	125.69	16.02	118.10	13.26	1.90	45.93	.06
IQ^a	105.53	10.85	109.39	9.22	−1.34	40.19	.19
Head motion	0.71	0.17	0.69	0.16	0.39	41.81	.70

Note. TDC = typically developing children; χ² = Pearson’s chi-square test; M = Male; F = Female; t = Welch’s t-test statistic.

IQ was available for 33 ADHD and 18 TDC.

Behavioral Data

Outcome measures are reported in Table 2. Importantly, task parameters were not significantly different between TDC and children with ADHD (all p > .20) after Bonferroni correction for multiple comparisons, except for the performance in the 0-Back control condition. Correlations between SSRT values and age were significant across the whole pool of participants (r_s = −.31, p = .02) and within the ADHD group, r = −.33, t (33) = −2.04, p = .049, but not within the TDC group (p = .09). Correlations between SSRT and IQ were not significant across the whole pool of participants, the ADHD group, or within the TDC group (all p > .54). Regarding the UP, correlations with age were not significant across the whole pool of participants, the ADHD group, or within the TDC group (all p > .15). Finally, correlations between UP and IQ were significant across the whole pool of participants (r_s = −.35, p = .01) and the TDC group (r_s = −.75, p > .001), but not within the ADHD group (p = .27).

Table 2.

Tasks Outcome Measures.

	ADHD (n = 36)		TDC (n = 20)		t	df	W	p
	M	SD	M	SD	t	df	W	p
Stop Signal Task^a
SSRT (ms)	292.00	91.37	283.23	73.89	0.39	46.76		.70
Correct stop trials %	48.12	15.22	45.62	9.63			355.5	.92
MRT	523.60	110.28	493.44	74.41			410	.20
SSD	270.19	217.10	210.22	62.08			382.5	.57
Omissions on Go trials	21.14	28.98	15.20	9.19			353.5	.95
N-Back^b
UP N0-N2	8.87	9.04	8.92	9.00			301.5	.80
N0	47.11	4.08	49.58	0.67			174.5	.01
N0-N2/N0	0.20	0.23	0.18	0.18			313.5	.97

Note. TDC = typically developing children; W = Wilcoxon rank-sum test statistic; SSRT = stop-signal reaction time; MRT = mean reaction time; SSD = Stop Signal Delay; UP = Updating Process from the N-Back; N0 = Performance at 0-Back; N2 = Performance at 2-Back; SST = Stop Signal Task.

SST performance scores were available for 55 participants (20 TDC and 35 ADHD).

N-Back performance scores were available for 53 participants (18 TDC and 35 ADHD).

Diffusion Measures: Between-Group Comparison

Results from the between-group comparison analysis are reported in Table 3. The whole-brain and the ROI TBSS analysis did not disclose any significant differences in FA or MD values between the ADHD and the TDC group (all p > .05).

Table 3.

Group Comparison Results and Correlation Analyses Between FA/MD and Behavioral Measures (Uncorrected p Values).

Region	Group comparison	Correlation SSRT			Correlation N0N2			Correlation N0
Region	Group comparison	Whole sample	ADHD	TDC	Whole sample	ADHD	TDC	Whole sample	ADHD	TDC
Whole brain	—	—	—	—	—	—	—	—	—	—
Cingulum	—	—	—	—	—	.020	—	—	—	—
CC	—	—	—	—	—	—	—	.038	—	—
SS	—	—	—	—	—	—	—	—	—	—
SLF	—	—	—	—	.044	.032	—	—	—	—

Note. FA = fractional anisotropy; MD = mean diffusivity; SSRT = stop-signal reaction time; TDC: typically developing children; CC = corpus callosum; SS = Sagittal Stratum; SLF = superior longitudinal fasciculus.

Correlation Analyses Between Diffusion Parameters and Inhibition—Working Memory Performance Measures

Results from the correlation analyses are reported in Table 3. The whole-brain TBSS correlation analysis did not show any significant association between FA or MD and task parameters from the SST or the N-Back scores, across the whole sample or the different groups (ADHD, TDC; all p > .05).

Regarding the ROI analysis, no statistically significant correlations were found between the SSRT and FA or MD in the whole sample or in the different groups separately, in any of the four preselected ROIs (all p > .05). The ROI correlation analysis with the N-Back scores disclosed four positive correlations approaching significance. A positive correlation between the UP score and MD in the SLF was found, across the whole sample (cluster size: 28; p = .044; Montreal Neurological Institute (MNI) coordinates: x = 37, y = −18, z = 33; Figures 3a and 4) and in the ADHD group (cluster size: 374; p = .032; MNI coordinates: x = 40, y = −40, z = 17; Figures 3b and 4). The UP score was positively correlated with MD in the cingulum in the ADHD group, with the largest cluster size located in the anterior part (cluster size: 116; p = .02; MNI coordinates: x = −6, y = 1, z = 34; Figures 3c and 4), and a positive correlation between the N0 score and FA in the CC was apparent across the whole sample (cluster size: 1,198; p = .038; MNI coordinates: x = −3, y = −35, z = 19; Figures 3d and 4). Nevertheless, none of these results survived Bonferroni correction (p < .017).

Figure 3.

Correlations between FA/MD and behavioral measures.

Figure 4.

Regions showing a significant correlation between FA or MD and performance parameters from the N-Back.

Discussion

Using TBSS, the present study aimed at investigating white matter microstructure in noncomorbid children with ADHD as compared with TDC, and at characterizing white matter tracks contributing to inhibition and working memory performance.

Correlation analyses between diffusion measures and behavioral indices revealed no correlation surviving Bonferroni correction for multiple comparisons at the whole-brain level or in the preselected ROIs. At trend level, a positive correlation between the UP score and MD was present in the SLF in the whole sample (p = .044) and in the ADHD group (p = .032). This finding would be consistent with Chiang et al. (2016) who reported an association between white matter microstructure in the SLF and spatial working memory performance in a group of children and adolescents with ADHD. In their study, GFA was negatively associated with total errors and strategy utilization in the Spatial Working Memory task from the Cambridge Neuropsychological Test Automated Battery (CANTAB) in the ADHD group. The SLF is the major fasciculus linking frontal and parietal lobes. It connects prefrontal and parietal regions that activate in tasks involving working memory. The association between SLF and verbal and/or spatial working memory has already been reported in healthy children, adolescents, and young adults (Østby, Tamnes, Fjell, & Walhovd, 2011; Peters et al., 2012; Vestergaard et al., 2011), but also in patients with recent-onset schizophrenia (Karlsgodt et al., 2008), patients who underwent prefrontal glioma resection (Kinoshita et al., 2016), diffuse traumatic brain injury (Palacios et al., 2011), and patients with multiple sclerosis (Bonzano, Pardini, Mancardi, Pizzorno, & Roccatagliata, 2009).

At trend level, a positive correlation between the UP score and MD in the cingulum was also present in the ADHD group (p = .020), predominantly in the anterior section. The cingulum connects prefrontal, parietal, and temporal regions and has been associated with working memory performance in TDC (Chiang et al., 2016), healthy middle-aged and elderly individuals (Charlton, Barrick, Lawes, Markus, & Morris, 2010), in individuals with temporal lobe epilepsy and left hippocampal sclerosis (Winston et al., 2013). Studies have suggested that both anterior and posterior regions of the cingulum are associated with working memory performance (Charlton et al., 2010; Short et al., 2013).

In our study, worse working performance appeared to be associated with higher MD. MD is a measure of the overall diffusivity in a particular voxel, regardless of fiber directionality. It is higher in areas where water diffuses more freely and lower where the movement of water molecules is more restricted, like in white matter. Variations in MD have been related to changes in intercellular space and tissue compactness (Beaulieu, 2002), but the precise neural correlates of altered MD measurements are uncertain. Nevertheless, the negative correlation between performance and MD indicates that these measurements may enlighten the mechanisms of a perturbed connectivity.

In addition to that, the N0 score from the N-Back, which is a measure of vigilance, was also associated with FA in the CC in the whole sample of participants at trend level (p = .038). A better performance in vigilance was associated with a higher FA. Voxels containing water that moves predominantly along a single direction have higher FA, and FA variations are associated with different factors, including myelination, axonal density, and fiber organization and integrity (Alba-Ferrara & de Erausquin, 2013). The CC is the largest white matter structure in the brain, connecting the hemispheres, and supports interhemispheric communication. Evidence exists for a role of the CC in sustained attention and vigilance (Rueckert & Levy, 1996). It is important to note that an atypical interhemispheric CC connection is the most common finding in studies comparing ADHD and TD participants (Aoki et al., 2017), and the involvement of the CC in the pathophysiology of ADHD has been evidenced by volumetric and functional studies. In a meta-analysis of ROI brain volume studies (Valera, Faraone, Murray, & Seidman, 2007), the splenium of the CC was one of the regions with the largest difference compared with controls. Reduced volume of the CC might affect the efficiency of interhemispheric communication, and therefore, cognitive functions that depend upon bilateral collaboration. McNally et al. (2010) found that a correlation between response speed and midbody circumference in boys (and isthmus circumference in girls) with ADHD were significantly correlated with response speed. In addition, Electroencephalography (EEG) studies have evidenced abnormal interhemispheric coherence in participants with ADHD (Hale et al., 2014).

No correlation was found between the SSRT and FA or MD in the whole sample or in the different groups separately, in any of the four preselected ROIs or the whole brain. Other studies have also failed to evidence a correlation between DTI-derived indices and response inhibition measures in participants with ADHD (Onnink et al., 2015) or typically developing individuals (Wu et al., 2016).

Finally, no between-group differences for FA/MD emerged in the whole brain or in our ROIs analysis. Nevertheless, it is important to note that there were no between-group differences in head motion. Our results are consistent with the ones of Aoki et al. (2017) who found that in the majority of studies in which there was no significant differences in head motion between groups, there were no significant differences in FA.

A limitation of the present study is that our sample included 16 participants with ADHD (44.4%) who were receiving medication (methylphenidate). In children and adolescents with ADHD, Lawrence et al. (2013) evidenced significantly higher MD and axial diffusivity (AD) in the forceps minor in medicated when compared with nonmedicated patients. On the other hand, de Luis-García et al. (2015) found significantly decreased MD in the CC and the corticospinal tract in children with ADHD under methylphenidate as compared with drug-naïve ADHD patients. Studies exploring the influence of treatment with methylphenidate on white matter microstructure of children with ADHD are thus contradictory, but we cannot rule out the possibility that medication influenced our results. On the other hand, one of the strengths of our study is the lack of significant between-group differences in head motion, as well as the inclusion of a quantitative measure of head motion in our model as confounding covariate. In addition, participants with ADHD had no psychiatric comorbidities, allowing the investigation of white matter abnormalities that would be specific to ADHD. Finally, our sample was characterized by a restrained, homogeneous age range, and age was entered as a confounding covariate in our model to control for potential age-related variability (Chen et al., 2016).

To conclude, consistent with a recent meta-analysis, no significant abnormalities in white matter diffusion indices was found in a sample of noncomorbid children with ADHD after correcting for head motion, providing support to the hypothesis that some previous findings were contaminated by motion artifacts. At trend level, an association between MD in the cingulum and the SLF and working memory in ADHD children, and an association between FA in the CC and vigilance in the whole sample were reported, in line with previous literature. Because these correlations did not survive Bonferroni correction for multiple comparisons (p < .017), replication studies are needed to assess whether they represent true effects or potential false positives. Taking into account previous findings, the CC might be important to characterize vigilance performance in ADHD, which is one of the main deficits in the disorder.

Footnotes

Acknowledgements

The authors thank all the children and their families for their participation.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by a Grant from the Belgian National Fund for Scientific Research (FNRS 3.4.516.08 .F). Ariadna Albajara Sáenz is supported by a grant from the Belgian Kids Foundation. Isabelle Massat and Ariadna Albajara Sáenz are supported by the Fonds National de la Recherche Scientifique (FNRS)-Belgium and the Fonds Erasme. Alison Mary was previously supported by a grant from the FNRS. Thomas Villemonteix was supported by a ULB fellowship. The authors declare no competing financial interest in relation to the work described.

ORCID iD

Philippe Peigneux

Author Biographies

Ariadna Albajara Sáenz is a PhD student in the Neuropsychology and Functional Neuroimaging Research Unit (UR2NF) at the Faculty of Psychology and Educational Sciences at the Université Libre de Bruxelles (ULB), under the supervision of Isabelle Massat and Philippe Peigneux.

Thomas Villemonteix is an assistant professor at the Laboratory of Neuropsychology and Psychopathology at Paris 8 Vincennes- Saint Denis University. He obtained his PhD in Psychology from Paris V-Paris Descartes University and held a postdoctorate fellowship in the Neuropsychology and Functional Neuroimaging Research Unit (UR2NF) at the Université Libre de Bruxelles (ULB).

Hichem Slama is a neuropsychologist at Erasme Hospital and Queen Fabiola Children’s University Hospital in Brussels and an assistant and scientific collaborator at the Faculty of Psychology and Educational Sciences at the Université Libre de Bruxelles. He obtained his PhD from the Faculty of Psychology and Educational Sciences at the Université Libre de Bruxelles (ULB).

Simon Baijot is a clinical neuropsychologist at the Queen Fabiola Children’s University Hospital in Brussels and a scientific collaborator at the Faculty of Psychology and Educational Sciences at the Université Libre de Bruxelles. He obtained his PhD from the Faculty of Psychology and Educational Sciences at the Université Libre de Bruxelles (ULB).

Alison Mary obtained her PhD from the Faculty of Psychology and Educational Sciences at the Université Libre de Bruxelles.

Danielle Balériaux is professor emeritus and senior consultant at the Department of Radiology at Erasme Hospital. During her career, she has been the head of the Neuroradiology department and the deputy head of the department of Radiology at Erasme Hospital (Université Libre de Bruxelles ULB). She received numerous honours for her pioneering work in MRI. She is presently the President of the Belgian Royal Academy of Medecine.

Thierry Metens is a lecturer at the Faculty of Medicine and the Polytechnic Institute at the Université Libre de Bruxelles (ULB). He obtained his PhD in Science (Physics) at the Université Libre de Bruxelles and is an MRI Physicist at the Radiology Department at Erasme Hospital since 1995.

Martin Kavec was a senior clinical research associate in Neuroradiology at the Department of Radiology, at the Clinics of Magnetic Resonance at Erasme Hospital (Université Libre de Bruxelles ULB).

Philippe Peigneux is tenure professor of Clinical Neuropsychology at the ULB in Belgium, where he is also the director of the Neuropsychology and Functional Neuroimaging Research Unit (UR2NF) affiliated at CRCN and at the ULB Neurosciences Institute (UNI). He obtained his PhD in Psychological Sciences from the University of Liège.

Isabelle Massat is a clinical Child Psychiatrist at Erasme Hospital and an FNRS qualified researcher at the Department of Experimental Neurology at the Université Libre de Bruxelles (ULB). She works in close collaboration with the Research Unit in Neuropsychology and Functional Neuroimaging (UR2NF) at the Faculty of Psychology and Educational Sciences at ULB. She obtained her PhD from the Faculty of Medicine at ULB.

References

Alba-Ferrara

de Erausquin

(2013). What does anisotropy measure? Insights from increased and decreased anisotropy in selective fiber tracts in schizophrenia. Frontiers in Integrative Neuroscience, 7, Article 9. doi:10.3389/fnint.2013.00009

American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders (4th ed., text rev.). Washington, DC: Author.

American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). Arlington, VA: American Psychiatric Publishing.

Aoki

Cortese

Castellanos

(2017). Research review: Diffusion tensor imaging studies of attention-deficit/hyperactivity disorder: Meta-analyses and reflections on head motion. Journal of Child Psychology and Psychiatry, 59, 193-202. doi:10.1111/jcpp.12778

Basser

(1995). Inferring microstructural features and the physiological state of tissues from diffusion-weighted images. NMR in Biomedicine, 8, 333-344. doi:10.1002/nbm.1940080707

Beaulieu

(2002). The basis of anisotropic water diffusion in the nervous system—A technical review. NMR in Biomedicine, 15, 435-455. doi:10.1002/nbm.782

Bechtel

Kobel

Penner

Specht

Klarhöfer

Scheffler

. . . Weber

(2012). Attention-deficit/hyperactivity disorder in childhood epilepsy: A neuropsychological and functional imaging study. Epilepsia, 53, 325-333. doi:10.1111/j.1528-1167.2011.03377

Bonzano

Pardini

Mancardi

Pizzorno

Roccatagliata

(2009). Structural connectivity influences brain activation during PVSAT in Multiple Sclerosis. NeuroImage, 44(1), 9-15. doi:10.1016/j.neuroimage.2008.08.015

Charlton

Barrick

Lawes

Markus

Morris

(2010). White matter pathways associated with working memory in normal aging. Cortex, 46, 474-489. doi:10.1016/j.cortex.2009.07.005

10.

Chen

Ouyang

Liao

Liu

. . . Gong

(2016). A systematic review and meta-analysis of tract-based spatial statistics studies regarding attention-deficit/hyperactivity disorder. Neuroscience & Biobehavioral Reviews, 68, 838-847. doi:10.1016/j.neubiorev.2016.07.022

11.

Chiang

Chen

Tsen

Gau

(2015). Altered white matter tract property related to impaired focused attention, sustained attention, cognitive impulsivity and vigilance in attention-deficit/ hyperactivity disorder. Journal of Psychiatry & Neuroscience, 40, 325-335. doi:10.1503/jpn.140106

12.

Chiang

Chen

Shang

Tseng

Gau

(2016). Different neural substrates for executive functions in youths with ADHD: A diffusion spectrum imaging tractography study. Psychological Medicine, 46, 1225-1238. doi:10.1017/s0033291715002767

13.

Congdon

Mumford

Cohen

Galvan

Canli

Poldrack

(2012). Measurement and reliability of response inhibition. Frontiers in Psychology, 3, Article 37. doi:10.3389/fpsyg.2012.00037

14.

Coxon

Van Impe

Wenderoth

Swinnen

(2012). Aging and inhibitory control of action: Cortico-subthalamic connection strength predicts stopping performance. Journal of Neuroscience, 32, 8401-8412. doi:10.1523/jneurosci.6360-11.2012

15.

de Luis-García

Cabús-Piñol

Imaz-Roncero

Argibay-Quiñones

Barrio-Arranz

Aja-Fernández

Alberola-López

(2015). Attention deficit/hyperactivity disorder and medication with stimulants in young children: A DTI study. Progress in Neuro-Psychopharmacology and Biological Psychiatry, 57, 176-184. doi:10.1016/j.pnpbp.2014.10.014

16.

Dennis

Francis

Cirino

Schachar

Barnes

Fletcher

(2009). Why IQ is not a covariate in cognitive studies of neurodevelopmental disorders. Journal of the International Neuropsychological Society, 15, 331-343. doi:10.1017/s1355617709090481

17.

DuPaul

(1998). ADHD Rating Scale-IV. New York, NY: Guilford Press.

18.

Endicott

Spitzer

R. L.

(1978). A diagnostic interview: The schedule for affective disorders and schizophrenia. Archives of General Psychiatry, 35, 837-844. doi:10.1001/archpsyc.1978.01770310043002

19.

Hale

Kane

Tung

Kaminsky

McGough

Hanada

Loo

(2014). Abnormal parietal brain function in ADHD: Replication and extension of previous EEG beta asymmetry findings. Frontiers in Psychiatry, 5, Article 87. doi:10.3389/fpsyt.2014.00087

20.

Karlsgodt

van Erp

Poldrack

Bearden

Nuechterlein

Cannon

(2008). Diffusion tensor imaging of the superior longitudinal fasciculus and working memory in recent-onset schizophrenia. Biological Psychiatry, 63, 512-518. doi:10.1016/j.biopsych.2007.06.017

21.

Keihaninejad

Ryan

Malone

Modat

Cash

Ridgway

. . . Ourselin

(2012). The importance of group-wise registration in tract based spatial statistics study of neurodegeneration: A simulation study in Alzheimer’s disease. PLoS ONE, 7(11), e45996. doi:10.1371/journal.pone.0045996

22.

King

Linke

Gass

Hennerici

Tost

Poupon

Wessa

(2012). Microstructure of a three-way anatomical network predicts individual differences in response inhibition: A tractography study. NeuroImage, 59, 1949-1959. doi:10.1016/j.neuroimage.2011.09.008

23.

Kinoshita

Nakajima

Shinohara

Miyashita

Tanaka

Okita

. . . Hayashi

(2016). Chronic spatial working memory deficit associated with the superior longitudinal fasciculus: A study using voxel-based lesion-symptom mapping and intraoperative direct stimulation in right prefrontal glioma surgery. Journal of Neurosurgery, 125, 1024-1032. doi:10.3171/2015.10.jns1591

24.

Lawrence

Levitt

Loo

Yee

O’Neill

. . . Narr

K. L.

(2013). White matter microstructure in subjects with attention-deficit/hyperactivity disorder and their siblings. Journal of the American Academy of Child & Adolescent Psychiatry, 52, 431-440.e4. doi:10.1016/j.jaac.2013.01.010

25.

Sun

Guo

Zang

Feng

Huang

. . . Gong

(2010). Increased fractional anisotropy in white matter of the right frontal region in children with attention-deficit/hyperactivity disorder: A diffusion tensor imaging study. Activitas Nervosa Superior Rediviva, 52, 193-199.

26.

Lipszyc

Schachar

(2010). Inhibitory control and psychopathology: A meta-analysis of studies using the stop signal task. Journal of the International Neuropsychological Society, 16, 1064-1076. doi:10.1017/s1355617710000895

27.

Logan

G. D.

(1994). On the ability to inhibit thought and action: A user’s guide to the stop-signal paradigm. In Dagenbach

Carr

T. H.

(Eds.), Inhibitory processes in attention, memory and language (pp. 189-239). San Diego, CA: Academic Press.

28.

Logan

G. D.

Van Zandt

Verbruggen

Wagenmakers

(2014). On the ability to inhibit thought and action: General and special theories of an act of control. Psychological Review, 121, 66-95. doi:10.1037/a0035230

29.

Madsen

Baaré

Vestergaard

Skimminge

Ejersbo

Ramsøy

. . . Jernigan

T. L.

(2010). Response inhibition is associated with white matter microstructure in children. Neuropsychologia, 48, 854-862. doi:10.1016/j.neuropsychologia.2009.11.001

30.

Massat

Slama

Kavec

Linotte

Mary

Baleriaux

. . . Peigneux

(2012). Working memory-related functional brain patterns in never medicated children with ADHD. PLoS ONE, 7(11), e49392. doi:10.1371/journal.pone.0049392

31.

Massat

Slama

Villemonteix

Mary

Baijot

Albajara Sáenz

. . . Peigneux

(2016). Hyperactivity in motor response inhibition networks in unmedicated children with attention deficit-hyperactivity disorder. The World Journal of Biological Psychiatry, 19, 101-111. doi:10.1080/15622975.2016.1237040

32.

McNally

Crocetti

Mahone

Denckla

Suskauer

Mostofsky

(2010). Corpus callosum segment circumference is associated with response control in children with attention-deficit hyperactivity disorder (ADHD). Journal of Child Neurology, 25, 453-462. doi:10.1177/0883073809350221

33.

Miller

Chapman

(2001). Misunderstanding analysis of covariance. Journal of Abnormal Psychology, 110, 40-48. doi:10.1037//0021-843x.110.1.40

34.

Nagy

Westerberg

Klingberg

(2004). Maturation of white matter is associated with the development of cognitive functions during childhood. Journal of Cognitive Neuroscience, 16, 1227-1233. doi:10.1162/0898929041920441

35.

Nichols

Holmes

(2002). Nonparametric permutation tests for functional neuroimaging: A primer with examples. Human Brain Mapping, 15(1), 1-25. doi:10.1002/hbm.1058

36.

Onnink

Zwiers

Hoogman

Mostert

Dammers

Kan

. . . Franke

(2015). Deviant white matter structure in adults with attention-deficit/hyperactivity disorder points to aberrant myelination and affects neuropsychological performance. Progress in Neuro-Psychopharmacology and Biological Psychiatry, 63, 14-22. doi:10.1016/j.pnpbp.2015.04.008

37.

Østby

Tamnes

Fjell

Walhovd

(2011). Morphometry and connectivity of the fronto-parietal verbal working memory network in development. Neuropsychologia, 49, 3854-3862. doi:10.1016/j.neuropsychologia.2011.10.001

38.

Owen

McMillan

Laird

Bullmore

(2005). N-back working memory paradigm: A meta-analysis of normative functional neuroimaging studies. Human Brain Mapping, 25, 46-59. doi:10.1002/hbm.20131

39.

Palacios

Fernandez-Espejo

Junque

Sanchez-Carrion

Roig

Tormos

, et al. (2011). Diffusion tensor imaging differences relate to memory deficits in diffuse traumatic brain injury. BMC Neurology, 11(1). doi: 10.1186/1471-2377-11-24

40.

Peters

Szeszko

Radua

Ikuta

Gruner

DeRosse

. . . Malhotra

A. K.

(2012). White matter development in adolescence: Diffusion tensor imaging and meta-analytic results. Schizophrenia Bulletin, 38, 1308-1317. doi:10.1093/schbul/sbs054

41.

Pierpaoli

Basser

(1996). Toward a quantitative assessment of diffusion anisotropy. Magnetic Resonance in Medicine, 36, 893-906. doi:10.1002/mrm.1910360612

42.

Pievsky

McGrath

(2017). The neurocognitive profile of attention-deficit/hyperactivity disorder: A review of meta-analyses. Archives of Clinical Neuropsychology, 33, 143-157. doi:10.1093/arclin/acx055

43.

Rossi

Moura

Mello

Souza

Muszkat

Bueno

(2015). Attentional profiles and white matter correlates in attention-deficit/hyperactivity disorder predominantly inattentive type. Frontiers in Psychiatry, 6, Article 122. doi:10.3389/fpsyt.2015.00122

44.

Rueckert

Levy

(1996). Further evidence that the callosum is involved in sustaining attention. Neuropsychologia, 34, 927-935. doi:10.1016/0028-3932(96)00009-7

45.

Schulze

Geary

Susmaras

Paliga

Maki

Little

(2011). Anatomical correlates of age-related working memory declines. Journal of Aging Research, 2011, Article 606871. doi:10.4061/2011/606871

46.

Shallice

Marzocchi

Coser

Del Savio

Meuter

Rumiati

(2002). Executive function profile of children with attention deficit hyperactivity disorder. Developmental Neuropsychology, 21, 43-71. doi:10.1207/s15326942dn2101_3

47.

Shang

Gau

Tseng

(2012). Disturbed microstructural integrity of the frontostriatal fiber pathways and executive dysfunction in children with attention deficit hyperactivity disorder. Psychological Medicine, 43, 1093-1107. doi:10.1017/s0033291712001869

48.

Short

Elison

Goldman

Styner

Connelly

. . . Gilmore

J. H.

(2013). Associations between white matter microstructure and infants’ working memory. NeuroImage, 64, 156-166. doi:10.1016/j.neuroimage.2012.09.021

49.

Smith

Jenkinson

Johansen-Berg

Rueckert

Nichols

Mackay

. . . Behrens

T. E.

(2006). Tract-based spatial statistics: Voxelwise analysis of multi-subject diffusion data. NeuroImage, 31, 1487-1505. doi:10.1016/j.neuroimage.2006.02.024

50.

Smith

Nichols

(2009). Threshold-free cluster enhancement: Addressing problems of smoothing, threshold dependence and localisation in cluster inference. NeuroImage, 44, 83-98. doi:10.1016/j.neuroimage.2008.03.061

51.

Takahashi

Iwamoto

Fukatsu

Naganawa

Iidaka

Ozaki

(2010). White matter microstructure of the cingulum and cerebellar peduncle is related to sustained attention and working memory: A diffusion tensor imaging study. Neuroscience Letters, 477, 72-76. doi:10.1016/j.neulet.2010.04.031

52.

Thomas

Sanders

Doust

Beller

Glasziou

(2015). Prevalence of attention-deficit/hyperactivity disorder: A systematic review and meta-analysis. Pediatrics, 135(4), e994-e1001. doi:10.1542/peds.2014-3482

53.

Valera

Faraone

Murray

Seidman

(2007). Meta-analysis of structural imaging findings in attention-deficit/hyperactivity disorder. Biological Psychiatry, 61, 1361-1369. doi:10.1016/j.biopsych.2006.06.011

54.

van Ewijk

Heslenfeld

Zwiers

Buitelaar

Oosterlaan

(2012). Diffusion tensor imaging in attention deficit/hyperactivity disorder: A systematic review and meta-analysis. Neuroscience & Biobehavioral Reviews, 36, 1093-1106. doi:10.1016/j.neubiorev.2012.01.003

55.

Vestergaard

Madsen

Baaré

Skimminge

Ejersbo

Ramsøy

. . . Jernigan

T. L.

(2011). White matter microstructure in superior longitudinal fasciculus associated with spatial working memory performance in children. Journal of Cognitive Neuroscience, 23, 2135-2146. doi:10.1162/jocn.2010.21592

56.

Wakana

Caprihan

Panzenboeck

Fallon

Perry

Gollub

. . . Mori

(2007). Reproducibility of quantitative tractography methods applied to cerebral white matter. NeuroImage, 36, 630-644. doi:10.1016/j.neuroimage.2007.02.049

57.

Wechsler

(1999). Wechsler Abbreviated Scale of Intelligence (WASI). San Antonio, TX: Psychological Corporation.

58.

Winston

Stretton

Sidhu

Symms

Thompson

Duncan

(2013). Structural correlates of impaired working memory in hippocampal sclerosis. Epilepsia, 54, 1143-1153. doi:10.1111/epi.12193

59.

Bralten

Cao

Hoogman

Zwiers

. . . Wang

Y. F.

(2016). White matter microstructural alterations in children with ADHD: Categorical and dimensional perspectives. Neuropsychopharmacology, 42, 572-580. doi:10.1038/npp.2016.223

60.

Yendiki

Koldewyn

Kakunoori

Kanwisher

Fischl

(2014). Spurious group differences due to head motion in a diffusion MRI study. NeuroImage, 88, 79-90. doi:10.1016/j.neuroimage.2013.11.027