Progress in Neuropsychopharmacology & Biological Psychiatry 104 (2021) 110044

Contents lists available at ScienceDirect

Progress in Neuropsychopharmacology
& Biological Psychiatry
journal homepage: www.elsevier.com/locate/pnp

Attention-deficit/hyperactivity disorder is characterized by a delay in T

subcortical maturation
⁎ ⁎ ⁎
Yanpei Wanga,b,c, , Chenyi Zuod, Qinfang Xua,b,c, , Lei Haob,c, Yuning Zhange,
a
Jiangsu Provincial Key Laboratory of Special Children's Impairment and Intervention, Nanjing Normal University of Special Education, Nanjing, China
b
State Key Laboratory of Cognitive Neuroscience and Learning, Beijing Normal University, Beijing, China
c
IDG/McGovern Institute for Brain Research, Beijing Normal University, Beijing, China
d
Faculty of Psychology, Beijing Normal University, Beijing, China
e
Centre for Innovation in Mental Health, University of Southampton, UK

A R T I C LE I N FO A B S T R A C T

Keywords: Although previous studies have found that ADHD is characterized by a delay in cortical maturation, it is not clear
ADHD whether this phenomenon was secondary to developmental trajectories in subcortical regions (caudate, pu-
Age tamen, pallidum, thalamus, hippocampus and amygdala). Using the ADHD-200 dataset, we estimated subcortical
Delay volumes in 339 individuals with ADHD and 568 typically developing controls. We defined the growth trajectory
Subcortical structures
of each subcortical structure, delineating a phase of childhood increase followed by an adolescent decrease in
Volume
subcortical volumes using a quadratic growth model. From these trajectories, the age of attaining peak sub-
cortical volumes was derived and used as an index of subcortical maturation. We found that subcortical struc-
tures (caudate, putamen, pallidum, thalamus, hippocampus and amygdala) followed curvilinear trajectories
similar to those reported in previous studies. The volumes of these subcortical structures in ADHD were also
delayed in the developmental trajectory, which suggested that ADHD may be characterized by a delay in sub-
cortical maturation. This delay may lead to a shift in which individuals with ADHD go through the process of
pruning the nerve connections that is part of the normal maturation process during adolescence. Further, we also
found that the asymmetric development of subcortical structures was abnormal in ADHD, which resulted from
the imbalance of the maturation delay of bilateral subcortical structures. The subcortical maturation delay may
play an important role in the pathophysiology of ADHD. Our findings provide new potential targets to in-
vestigate the pathophysiology of ADHD.

1. Introduction delay in the growth of the brain regions that are not functioning
properly in ADHD individuals, and they found that ADHD children and
The incidence of attention-deficit/hyperactivity disorder (ADHD) their younger but typical developing peers have similar brain activity at
among school-age children is from 5% to 15% (Asherson et al., 2016; rest and at cognitive task (El-Sayed et al., 2003; Wang et al., 2019;
Shaw et al., 2007), and ADHD have been a most common neuro-de- Wang et al., 2020), which is consistent with the lag in cortical devel-
velopmental disorder in childhood. This syndrome seriously affects opment (Shaw et al., 2007). However, others have reported a quanti-
school performance. (Castellanos and Tannock, 2002). Many factors tative and different neurophysiology with some highly unusual findings
(e.g. genetic factors, maternal stress and perinatal trauma) increase the in the study of unique electroencephalogram (EEG) functional imaging,
risk for ADHD (Child NSCO, 2008), however, there is debate as to which is more similar to a typical development deviation (Hobbs et al.,
whether there is a fundamental deviation from the typical develop- 2007).
mental growth pattern or a growth delay in areas of the brain that are The cerebral cortex grows gradually from birth for a period of time
not functioning properly. As early as 1997, Barkley proposed that and finally reaches a peak between 7 and 11 years old and then goes
ADHD is caused by an abnormal prefrontal cortex and its connections to through a normal period of decline because of the pruning of less ef-
the striatum, and this abnormality leads to serious changes in inhibitory fective neural connections (Child NSCO, 2008). The brain grows at
(Barkley, 1997). Subsequently, a series of studies found that there is a different rates in different children, but we do not understand the

⁎
Corresponding authors at: Jiangsu Provincial Key Laboratory of Special Children's Impairment and Intervention, Nanjing Normal University of Special Education,
Nanjing, China.
E-mail addresses: [email protected] (Y. Wang), [email protected] (Q. Xu), [email protected] (Y. Zhang).

https://doi.org/10.1016/j.pnpbp.2020.110044
Received 19 February 2020; Received in revised form 12 June 2020; Accepted 13 July 2020
Available online 18 July 2020
0278-5846/ © 2020 Elsevier Inc. All rights reserved.
Y. Wang, et al. Progress in Neuropsychopharmacology & Biological Psychiatry 104 (2021) 110044

significance of this unique growth. The difference between a delay in sensorimotor cortex (Mostofsky et al., 2006). Consistent with these
growth and the completely different brain developmental patterns in functional defects is the finding of structural compromise(Ellison-
ADHD children could contribute to clarify the bases of ADHD and even Wright et al., 2008; Dickstein et al., 2006; Valera et al., 2007). Reduced
to develop better strategies for the treatment of children with ADHD. total striatum volumes are often, but not always, reported in preschool-
Previous studies have found that ADHD disorder was characterized by a age children.
delay in cortical maturation, especially in the prefrontal cortex (Shaw The thalamus plays an important role in the pathogenesis of ADHD.
et al., 2007; Shaw et al., 2013; Shaw et al., 2012). Compared with ty- The thalamus projects to the frontal and parietal association cortices
pical developing (TD) children, the age of reaching peak cortical (Romanski et al., 1997), two thalamocortical pathways that support
thickness in the ADHD individuals was delayed. The cortical thickness attention processes. These circuits are important for stimulus seeking
reached peak at 7.5 years of age in TD children, while delayed to and distinguishing environment-related and environment-unrelated
10.5 years in ADHD children. There were significant differences be- somatosensory stimuli (Wester et al., 2001). Within this overall thala-
tween the two groups in the cognitive and emotional regulation areas of mocortical network, the thalamus integrates and coordinates responses
the cerebral cortex, especially the frontal lobe, where there was a 5-year to auditory and visual stimuli (Kastner et al., 2004). Morphological
difference in reaching peak thickness. Brain areas such as the insula and disturbances in the thalamus in youths with ADHD therefore may
striatum are affected (Frodl and Skokauskas, 2012; Nakao et al., 2011) contribute to differences in allocating and directing attentional re-
although the changes in the prefrontal cortex might be the most pro- sources toward salient stimuli. The thalamus also connects with limbic
minent (Shaw et al., 2012; Sonuga-Barke and Halperin, 2010). One structures, including the amygdala (Liddell et al., 2005). Some studies
study also reported that the degree of delay in prefrontal lobe ma- have previously reported morphological disturbances in the basolateral
turation was related to the severity of attention deficit symptoms in complex of the amygdala (Plessen et al., 2006), and others have re-
adults (Shaw et al., 2013). ported abnormal amygdala activation (Marsh et al., 2008) in ADHD
Contemporary theories of the pathogenesis of ADHD conceptualize group. The findings about the amygdala seem more consistent, with
symptoms of inattention, hyperactivity, and impulsivity as con- most studies reporting decreased volumes, (Bonath et al., 2018; Frodl
sequences of dysfunction in the cortical-subcortical loops (Eickhoff and Skokauskas, 2012; Hoogman et al., 2017) or no difference (Amico
et al., 2009; Grefkes et al., 2008). Researchers found different con- et al., 2011; Perlov et al., 2008) in ADHD group compared with healthy
nectivity between the cortex and subcortex, such as the cortico-striato- controls. The hippocampus likely subserves many functions in attention
thalamo-cortical (CSTC) loops, the prefrontal-striatal model, the tha- and cognition, and their dysfunction is one of the main characteristics
lamocortical loop and the fronto-striatal-cerebellar circuit, which sub- of ADHD. (Castellanos and Tannock, 2002; Pliszka, 2000). Previous
serve among other processes, executive function and effortful control studies have reported unclear results for the volume changes of hip-
(Rowe et al., 2005). In addition, the animal model also showed that the pocampus in the ADHD individuals. Scholars have observed increased
potentiated limbic (hippocampus and amygdala) loop together with the volume (Plessen et al., 2006), decreased volume (Bonath et al., 2018;
tuned-down cortical loop may play a key role in controlling ADHD-like Hoogman et al., 2017; Posner et al., 2014) or similar volume (Amico
behavioral symptoms in rats (Castellanos and Aoki, 2016). Abnormal et al., 2011; Perlov et al., 2008). Inconsistent results that include in-
activation in these networks is thought to produce hyperarousal and an creased volume, decreased volume or no difference in subcortical
inadequate suppression of both irrelevant sensory inputs and premature structures may be related to different age ranges or different meth-
behavioral responses, which in turn manifest as the cardinal symptoms odologies. Shaw et al. found the cortical thickness reached peak value
of inattention, hyperactivity, and impulsivity that define ADHD in TD children was 7.5 years old, while delayed to 10.5 years of age in
(Dickstein et al., 2006; McFarland and Haber, 2002; Rowe et al., 2005). the ADHD-diagnosed children. Raznahan et al. (2014) found that sub-
Animal studies have shown that CSTC loops convey information from cortical structures also followed curvilinear trajectories akin to those
the cortex to the basal ganglia, then to the thalamic nuclei, and then reported for the cortical sheet(Raznahan et al., 2014). Therefore, the
back to the cortex (Alexander et al., 1986). In principle, disturbances in different age ranges may result in different volumetric changes.
any of the structures along these CSTC pathways could produce ab- Normal variation and functional specialization lead to structural,
normal information processing and ultimately produce the symptoms of functional and behavioral asymmetries, which have evolved over the
ADHD (Dickstein et al., 2006; Perlov et al., 2008). The neurobiological course of the lifespan and are thought to have originated from devel-
model of ADHD mainly focuses on the disturbance in the relationship opmental, genetic, empirical and pathological factors (Toga and
between the frontal lobe and its subcortical structure, which leads to Thompson, 2003). Attention itself can lead to functional asymmetries in
executive dysfunction, motor deficit and difficult regulation of atten- healthy individuals. In ADHD, early pathological studies suggest that
tion, motivation and affect, which in turn leads to the behavioral unilateral right hemispheric dysfunction plays a key role (Stefanatos
symptoms of ADHD. and Wasserstein, 2001). Previous studies have examined many aspects
The basal ganglia (including caudate, putamen and pallidum) is of brain structural symmetry in ADHD (Hale et al., 2015; Shaw et al.,
related to the pathophysiology of ADHD, but little is known about its 2009). One of these studies observed decreased asymmetry in ADHD,
development. The basal ganglia are the core component of the closely compared with homologous cortical regions (Shaw et al., 2009). In a
connected “loops” between the cerebral cortex and the thalamus, which fMRI study of ADHD, increased or atypical functional asymmetries and
supports many cognitive processes that are damaged in ADHD. abnormal inter-hemispheric processing have been found (Fassbender
Dysfunction of the ventral striatum (caudate and putamen), pallidum and Schweitzer, 2006). The maturation rates of asymmetry between
and limbic cortex (hippocampal and amygdala) is mainly associated hemispheres may be a key detail in understanding the changes in brain
with the abnormal reward processing found in ADHD (Sonuga-Barke, structure throughout the lifespan. Therefore, we tested asymmetry in
2005). Moreover, the ENIGMA working group also carried out a meta- volumes.
analysis base on the structural alterations of subcortex and found a In general, there is controversy about the nature of brain develop-
smaller caudate, putamen and intracranial volumes in ADHD in- mental disorder, which is the basis of ADHD. Our question is that
dividuals than in a control group(Hoogman et al., 2017). Problems in whether the ADHD disorder is characterized by a delay in subcortical
executive function, for example, cognitive control and working maturation. We thought that the growth pattern of the brain in ADHD-
memory, are related to the abnormality in the circuit connecting the diagnosed children is essentially different from that of children with
lateral prefrontal cortex and the head of caudate and anterior putamen other neurological disorders (such as autism and intelligence dis-
(Hart et al., 2013). Motor planning and control problems are another ability), and the developmental pattern of these disorders is different
feature of ADHD, that may be due to the interruption of the connection from those with normal development. On the contrary, we hypothesize
between the posterior/caudal areas of the basal ganglia and the the maturation of brain structures in TD children and in children with

2
Y. Wang, et al. Progress in Neuropsychopharmacology & Biological Psychiatry 104 (2021) 110044

ADHD occurs in a highly similar manner, but children with ADHD have effects of age and diagnosis on the subcortical volumes from the
significant delays. This delay can also cause ADHD patients to undergo quadratic model to the linear model, including the main effects of di-
pruning of neural connections during adolescence as part of a normal agnosis, age and the interaction of diagnosis and age (Narvacan et al.,
maturation process. Previous studies hypothesized that ADHD was re- 2017). Each subcortical volume was fitted to (i) Quadratic model:
lated to globally decreased brain volumes compared with TD controls. If Subcortical volume = Intercept + A(Diagnosis) + B
the sample tended to be younger, the relationship between age and (Age) + C(Age2) + D(Age × Diagnosis) + E
volume showed a positive linear trajectory; however, when sample (Age2 × Diagnosis) + residual error; or (ii) Linear model: Subcortical
tended to be older, the relationship was negative with the sample volume = Intercept + A(Diagnosis) + B(Age) + C(Age × Diag-
evenly distributed on both sides of peak, and the linear relationship nosis) + residual error. If the age and diagnosis interaction was not
disappeared. Thus, to test this hypothesis, we tested for the existence of significant, each ERC measure was fit with only diagnosis and age main
diagnosis-related differences in age-dependent subcortical volume effects. There were a total of four fits for each measure to choose in the
changes from childhood to adulthood in the ADHD-200 Consortium. We model. If a measure fit a quadratic model, we calculated the age at the
hypothesized that the ADHD disorder is characterized by a delay in peak from the first derivative. Statistical significance was p < .05,
subcortical maturation, therefore, the diagnosis × age interaction ef- adjusting for the number of comparisons. Akaike Information Criterion
fects would be specific for distinct subcortical structures, and depend on (AIC) and Bayesian Information Criterion (BIC) also was used to com-
the age of the volume peak. In addition, as studies have reported al- pare models and select the most probable age trajectory from multiple
terations in the normal symmetry of the subcortical structure (Hale fits within the same structure (Tamnes et al., 2013; Wagenmakers and
et al., 2015; Shaw et al., 2009; Toga and Thompson, 2003; Wang et al., Farrell, 2004). AIC and BIC take into consideration the likelihood of the
2018), we tested for asymmetry in volumes as well. model and the number of parameters being fitted for the dataset, while
penalizing models with more parameters. The fit with the lowest AIC or
2. Materials and methods BIC was chosen as the best fit model for each structure, the results were
shown in Table S1 and Table S2.
2.1. Participants and data acquisition We used Bonferroni correction for multiple comparisons. In order to
optimally balance the Type-I and Type-II errors, we used the Simple
The data we used in this study were publicly available from the Interactive Statistical Analysis Bonferroni tool (http://www.
ADHD-200 Consortium (http://fcon_1000.projects.nitrc.org/indi/ quantitativeskills.com/sisa/calculations/bonfer.htm) to consider the
adhd200/). The ADHD-200 dataset contains functional and anato- correlation beween the dependent variables. Using a Bonferroni cor-
mical MRI data provided by eight institutions. Each cohort was ap- rection that treats the variables as independent variables would lead to
proved by the research ethics review boards of each institution. All a too stringent correction, as the dependent variables are not obtained
participants or their legal guardians signed informed consent before in independent subgroups. The mean correlation of subcortical volumes
participating. was r = 0.66, and the corresponding corrected alpha value was 0.021.
A total of 67 participants were excluded due to the lack of diag- Subcortical asymmetries showed a mean correlation coefficient of
nostic information (26 participants) and image quality (41 partici- r = 0.06, and the significant level of alpha = 0.005 was equivalent to a
pants). We had a strict quality control for each structural T1-weighted corrected p = .05.
brain MRI data. We carefully conducted visual inspections by two ex-
perienced raters, the detailed protocol was similar to the that used in 3. Results
the Human Connectome Project (Marcus et al., 2013). If both raters
considered that the image have a better than fair quality, we would 3.1. Subcortical volumes
used the image in the subsequent analysis. Finally, we selected 906
participants aged 7–22 years for subsequent analysis, including 338 The best fit models for each subcortical gray matter volume are
individuals with ADHD (male n = 263; female n = 75, mean shown in Table 1, Table 2 and Fig. 2. In these models, the main effects
age = 11.73 ± 3.06) and 568 TD individuals (male n = 296; female of diagnosis were observed significantly in the bilateral caudate(left:
n = 272, mean age = 12.17 ± 3.37). The participant information and β = −1.05, t = −2.29, p = .022; right: β = −1.11, t = −2.39,
the detailed distributions of participants are shown in Table S1, and the p = .017), pallidum(left: β = −0.51, t = −3.46, p = .001; right:
detailed phenotype information for each subject can be found on the β = −0.46, t = −3.21, p = .001) and amygdala(left: β = −0.13,
website for the ADHD-200 Consortium (http://fcon_1000.projects.nitrc. t = −3.08, p = .002; right: β = −0.15, t = −3.73, p < .001),
org/indi/adhd200/). showing that ADHD individuals had smaller volumes than TD in-
dividuals. The bilateral caudate, putamen, pallidum, thalamus and
2.2. MRI processing hippocampus fit the quadratic models, while the bilateral amygdala fit
the linear increased models. The bilateral caudate (left:
To process structural T1-weighted brain MRI data, we used β = −5.71 × 10−3, t = −2.06, p = .039; right: β = −6.10 × 10−3,
“volBrain”, which is an advanced pipeline that automatically provides t = −2.17, p = .030) and pallidum (left: β = −2.39 × 10−3,
volumetric information on the brain MR images at different scales t = −2.67, p = .008; right: β = −2.22 × 10−3, t = −2.57, p = .010)
(Manjón and Coupe, 2016). The system provides the volumes/seg- showed significant interaction effects between age and diagnosis. Fur-
mentations and structural asymmetries in subcortical gray matter ther analysis found that the bilateral caudate (age at peak volume: left:
structures (putamen, caudate, pallidus, thalamus, hippocampus and 12.53,right:12.50) and pallidum (age at peak volume: left:13.52,-
amygdala -, Fig. 1). The asymmetry index is calculated as the difference right:12.95) fit the quadratic models in ADHD individuals, while they
between the right and left volumes divided by their mean (in percent) fit the linear decreased models in TD controls. The bilateral putamen
(Shaw et al., 2009; Wang et al., 2018). (age at peak volume: ADHD: left:14.39,right:12.95; TD: left:14.25,-
Asymmetry = (left volumes-right volumes)/mean (left volumes, right:10.92), thalamus (age at peak volume: ADHD:left:13.67,-
right volumes) × 100%. right:12.93; TD:left:11.33,right:9.91) and hippocampus (age at peak
volume: ADHD: left:19.17,right:16.90; TD: left:16.75,right:15.53)
2.3. Statistical analysis showed no significant interaction effects between age and diagnosis,
and both ADHD and TD individuals fit the quadratic models, but the
All statistical analyses were conducted using SPSS 23.0(SPSS, peak volumes were reached at different ages. The bilateral amygdala
Chicago, IL). We used linear and non-linear regressions to estimate the (left: β = 8.01 × 10−3, t = 2.38, p = .018; right: β = 1.05 × 10−2,

3
Y. Wang, et al. Progress in Neuropsychopharmacology & Biological Psychiatry 104 (2021) 110044

Fig. 1. 3D view of subcortical segmentations.

t = 3.14, p = .002) show significant interaction effects between age understand specific patterns of many neurological diseases (Manjón and
and diagnosis. Further analysis found that, compared TD individuals Coupe, 2016). We found contributions of diagnosis to volume and
(left: β = 3.71 × 10−3, t = 2.11, p = .035; right: β = 3.65 × 10−3, asymmetry, including the volume of the bilateral caudate, pallidum and
t = 2.03, p = .043), ADHD individuals (left: β = 1.17 × 10−2, amygdala and the asymmetry of the putamen and thalamus. The bi-
t = 3.75, p < .001; right: β = 1.42 × 10−2, t = 4.69, p < .001) lateral volume of the caudate, putamen, pallidum, thalamus and hip-
showed stronger linear growth. pocampus and the asymmetry of the hippocampus fit quadratic models,
while only the bilateral volume of the amygdala and the asymmetry of
the thalamus fit linear models. Remarkably, the observed age relations
3.2. Subcoritcal asymmetry
depended on the diagnosis of individuals: significant age×diagnosis
interaction effects were observed on volumes of the bilateral caudate,
All the results in subcortical asymmetry are detailed in Table 3,
pallidum and amygdala and the asymmetry of the hippocampus. Spe-
Table 4 and Fig. 3. We observed a significant diagnosis main effect in
cifically, TD individuals show a linear decrease in the bilateral caudate
the asymmetry of the putamen (β = 0.43, t = 2.85, p = .004) and
and pallidum, while ADHD individuals fit quadratic models. The
thalamus (β = 0.43, t = 2.29, p = .022). The asymmetry of the tha-
amygdala gray matter volume showed a stronger increase in ADHD
lamus showed a significant decrease with age (β = −0.09, t = −3.21,
individuals than in TD individuals. In regard to hippocampus asym-
p = .001) but no significant interaction. The asymmetry of the hippo-
metry, we found that only ADHD individuals fit a quadratic model, and
campus showed a significant interaction (β = −0.06, t = −1.97,
TD individuals did not fit any models. Although we did not observe an
p = .049). Further analysis found that individuals with ADHD show a
interaction between age and diagnosis in the bilateral putamen, tha-
significant quadratic model (age at peak volume:13.59) but no age ef-
lamus and hippocampus, we found that ADHD individuals reached the
fect in TD controls.
peak volume value earlier than TD indviduals.
The basal ganglia include a group of nuclei (caudate,putamen and
4. Discussion pallidum) involved in cognitive, emotional and motor behavior
(Alexander et al., 1986). In addition to the prefrontal cortex, the cau-
In this study, we investigated the effects of age on subcortical date and its related circuits are also associated with ADHD (Pontius,
structures vital for cognitive and emotional adaptation to daily life from 1973). The striatum (caudate and putamen) is the entry point to the
childhood to adulthood, while explicitly testing differential relations in basal ganglia, and anomalies of both structures are reported in ADHD.
TD and ADHD individuals. We used an automated reliable quantitative In voxel-based studies, the volume loss of the right striatum is em-
analysis tool “volBrain”, which is a novel and fully automatic pipeline phasized as a major feature of ADHD, which is “unknown”, that is to
for volumetric brain analysis based on multiatlas label fusion tech- say, it is not limited to preplanned analysis (Nakao et al., 2011). In our
nology, which allowed us to analyze brain development and to

Table 1
Fitting parameters for subcortical structures versus age within each hemisphere.
Parameter(S.E.)

Measures HS Best fit model Intercepta Diagnosis Age×10−1 Age2 Age×sex Age2 × sex
×10−3 ×10−2 ×10−3

Caudate L Quadratic 3.86(0.26) −1.05(0.46) n.s. n.s. 15.59(7.29) −5.71(2.77)

R Quadratic 3.99(0.27) −1.11(0.46) n.s. n.s. 16.64(7.39) −6.10(2.81)
Putamen L Quadratic 4.02 (0.23) n.s. 1.45 (0.37) −5.87(1.37) – –
R Quadratic 3.96(0.23) n.s. 1.51 (0.36) −6.16(1.34) – –
Pallidum L Quadratic 1.53(0.09) −0.51(0.15) n.s. n.s. 7.61(2.36) −2.39(0.90)
R Quadratic 1.49(0.08) −0.46(0.14) n.s. n.s. 6.82(2.27) −2.22(0.86)
Thalamus L Quadratic 5.71 (0.26) n.s. 1.43 (0.41) −5.76(1.51) – –
R Quadratic 5.83 (0.24) n.s. 1.14(0.38) −4.94(1.42) – –
Hippocampus L Quadratic 2.70(0.19) n.s. 1.38(0. 30) −4.21(1.11) – –
R Quadratic 2.52(0.19) n.s. 1.80(0.30) −5.66(1.13) – –
Amygdala L Linear 0.81 (0.02) −0.13 (0.04) 0.04(0.02) – 0.80 (0.34) –
R Linear 0.82(0.02) −0.15 (0.04) n.s. – 1.05(0.33) –

n.s. = not significant;– = not applicable; HS = Hemisphere; L = left; R = Right.

a
Intept is the extrapolated value at age zero.

4
Y. Wang, et al. Progress in Neuropsychopharmacology & Biological Psychiatry 104 (2021) 110044

Table 2
Fitting parameters for subcortical volumes versus age within each hemisphere in TD and ADHD individuals.
Parameter(S.E.)

Structure HS TD ADHD

Best fit model Intercepta Age Age2 Best fit model Intercepta Age Age2
(×10−1) (×10−3) (×10−1) (×10−3)

Caudate L Linear 4.30(0.07) −0.18(0.06) – Quadratic 2.81(0.37) 2.08(0.60) −8.30(2.31)

R Linear 4.34(0.07) −0.18(0.06) – Quadratic 2.88(0.38) 2.05(0.61) −8.19(2.34)
Putamen L Quadratic 4.33(0.30) 1.07(0.46) −4.89(1.69) Quadratic 3.69(0.39) 1.70(0.62) −5.91(2.40)
R Quadratic 4.26(0.29) 1.14(0.45) −5.20(1.66) Quadratic 3.66(0.38) 1.74(0.61) −6.10(2.34)
Pallidum L Linear 1.50(0.02) −0.11(0.02) – Quadratic 1.02(0.12) 0.60(0.20) −2.22(0.75)
R Linear 1.47(0.02) −0.11(0.02) – Quadratic 1.03(0.11) 0.54(0.18) −2.09(0.68)
Thalamus L Quadratic 5.99(0.34) 1.06(0.53) −4.70(1.96) Quadratic 5.36(0.39) 1.85(0.62) −6.75(2.40)
R Quadratic 6.14(0.32) 0.72(0.50) −3.63(1.84) Quadratic 5.40(0.36) 1.74(0.58) −6.72(2.23)
Hippocampus L Quadratic 2.76(0.24) 1.34(0.36) −4.29(1.34) Quadratic 2.73(0.33) 1.15(0.53) −2.82(2.06)
R Quadratic 2.71(0.23) 1.56(0.36) −5.02(1.32) Quadratic 2.29(0.36) 1.99(0.57) −5.89(2.21)
Amygdala L Linear 0.81(0.02) 0.04(0.02) – Linear 0.68(0.04) 0.12(0.03) –
R Linear 0.82(0.02) 0.04(0.02) – Linear 0.66(0.04) 0.14(0.03) –

– = not applicable; n.s. = non-significant; HS = Hemisphere; L = left; R = Right.

a
Intercept is the extrapolated value at age zero.

study, we found a smaller volume in the bilateral caudate. Several The putamen is an area associated with primary and supplementary
groups also observed a decrease in the caudate (Castellanos et al., motor areas, which may lead to the motor symptoms of ADHD, and the
2002), but some studies did not find differences(Pineda et al., 2002), study of the putamen has produced equally ambiguous results.
even though some studies found a larger caudate in ADHD children Consistent with our data, the researchers examined the putamen vo-
than in TD children (Garrett et al., 2008). Mataro et al. (1997) found lume as an area of interest but found no significant differences
that decreased volumes were detected in the group with ADHD for (Wellington et al., 2006). However, it was found that there was a sig-
those younger than 16, but not for those older than 16 (Mataro et al., nificant reversal of asymmetry between groups. Compared with TD
1997). The different results in previous studies may be due to the use of controls, the left putamen of ADHD individuals was smaller than the
different age goup samples. Indeed, we found that the caudate followed right putamen, which was consistent the results of a meta-analysis
different developmental trajectories in ADHD and TD individuals. We study and a longitudinal study (Pretus et al., 2017). The finding of this
found that the caudate volume in TD controls decreased with age, symmetry reversal indirectly supports the finding of decreased blood
which is consistent with the result of a previous study (Wierenga et al., flow to the putamen, especially to the left side (Teicher et al., 2002)
2014), in which they investigated the typical development of the basal (Teicher et al., 2002). Teicher et al. (2000) found that measures of
ganglia from age 7 to 24, which was similar to that of our sample, and hyperactivity and inattention could be accounted for by deficits in
they observed the same result. Meanwhile, the caudate volume in blood flow to the putamen alone, and the deficits in blood flow and
ADHD individuals does not undergo a spatially homogenous linear behavior were both attenuated by methylphenidate (Teicher et al.,
change with development and follows a highly dynamic and regionally 2000). Some studies have also found that damage to the putamen
heterogenous “inverted-U" trajectory, that is, first increased with age, caused by stroke or traumatic brain injury is associated with the de-
reached peak at approximately 12.5 years old, and then decreased with velopment of attention deficit disorder (Max et al., 2002). Therefore,
age. We hypothesized that TD controls would also follow a “inverted-U" the reversal of putamen asymmetry may lead to the motor symptoms of
trajectory, however, if the peak occurs earlier, our sample did not ADHD. These findings, together with the putamen's function, suggest
capture it. For example, Carrey et al. (2012) found a different pattern of that the possibility of putamen abnormalities is associated with the
caudate volume abnormalities across narrow age clusters(Carrey et al., impulsivity and hyperactivity expressed in this group of children. We
2012); that is, children with ADHD had smaller total caudate volumes did not found a diagnosis×age interaction, but we did find that the
than controls in younger participants (5.9–7.3), but no diagnosis effect bilateral putamen volume fit an “inverted-U" developmental trajectory.
was found in older participants (7.4–10.8). However, they did not pay In further analysis, we also found that the volume of the putamen fit a
attention to the developmental trajectory of the caudate, perhaps be- quadratic developmental trajectory both in ADHD and TD individuals,
cause the younger participants did not reach the peak of caudate vo- reaching peak volumes at 14.25 and 10.92 years in TD, and at 14.39
lume due to the developmental delay of ADHD children. To test this and 12.95 years in ADHD individuals in the left and right hemispheres,
hypothesis, we divided our data into two sections between peaks and respectively. From the trajectory, we found that the matruation delay in
obtained the same results. Our findings were consistent with the largest ADHD individuals was observed only in the right putamen. Ad-
longitudinal MRI study of developmental trajectories in ADHD ditionally, the maturation delay of the right putamen may be the reason
(Castellanos et al., 2002), in which the caudalte was smaller in pre- for the abnormality of putamen asymmetry.
pubertal children with ADHD than in healthy controls. Our and others' The caudate and putamen play key roles in several basal ganglia-
findings (Mahone et al., 2011) extend the results of previous studies and thalamus-cortical circuits involved in motion control and learning, as
show that decreased caudate volumes can be detected in ADHD chil- well as in the selection and activation of cognitive, executive and
dren as early as ages 4 or 5. In addition, these findings suggest that the emotional procedures. Several of these processes (e.g., motor function,
volume difference of the caudate begins to disappear before adoles- reward processing, cognition and attention control) are damaged in
cence, that is, at the age of 8. It is worth noting that ADHD symptoms ADHD. Previous studies have found that striatum is smaller or different
(mainly hyperactivity/impulse symptoms) decrease with age (Hinshaw in ADHD in dividuals than in control individuals. In addition, different
et al., 2006). Hyperactive/impulsive symptoms in younger children developmental trajectories in participants with ADHD and in control
may reflect the disruption of the basal ganglia during a developmental individuals could provide a potential explanation for the results of
period while cortical areas (especially the prefrontal cortex) are not previous studies, indicating changes in striatum volume with age. In
fully functional (Soliva et al., 2010). ADHD, the delay of the developmental trajectories of the caudate and

5
Y. Wang, et al. Progress in Neuropsychopharmacology & Biological Psychiatry 104 (2021) 110044

Fig. 2. The age trajectories of left (upper layer graph)and right (lower layer graph) subcortical volumes, including the (A) caudate, (B) putamen, (C) pallidum, (D)
thalamus, (E) hippocampus, (F) amygdala in TD (red color) and ADHD individuals (blue color). (For interpretation of the references to color in this figure legend, the
reader is referred to the web version of this article.)

6
Y. Wang, et al. Progress in Neuropsychopharmacology & Biological Psychiatry 104 (2021) 110044

Table 3
Fitting parameters for asymmetry of subcortical structures versus age.
Parameter(S.E.)

Structure Best fit model Intepta Diagnosis Age Age2 Age×sex Age2 × sex
×10−1 ×10−2 ×10−2

Caudate None – – – – – –
Putamen None n.s. 4.26(1.50) n.s. – – –
Pallidum None – – – – – –
Thalamus Linear n.s. 4.25(1.85) −8.83(2.75) – – –
Hippocampus Quadratic n.s. n.s. n.s. n.s. 1.66 (0.85) −6.32(3.21)
Amygdala None – – – – – –

n.s. = not significant;– = not applicable.

a
Intercept is the extrapolated value at age zero.

putamen may be the potential cause of the results observed in this environment-related and environment-unrelated somatosensory stimuli
study. (Wester et al., 2001). Within this overall thalamocortical network, the
The pallidum was found to be significantly smaller in ADHD chil- thalamus integrates and coordinates responses to auditory and visual
dren than in TD children, a result that is also similar to that of a pre- stimuli (Kastner et al., 2004; McAlonan et al., 2008). Morphological
vious study (Rosch et al., 2018). The study, which was based on ADHD disturbances in the thalamus in youths with ADHD therefore may
animal models, found that the early pallidum delay stabilized before contribute to differences in allocating and directing attentional re-
6 weeks (human equivalent of 7–9 years old) (Hsu et al., 2010), and a sources toward salient stimuli. The thalamus also connects with limbic
cross-section study of subcortical volume reduction in ADHD showed structures, including the amygdala (Liddell et al., 2005). Therefore, the
that subcortical differences decreased during development (Hoogman interference in the thalamolimbic network could affect emotional
et al., 2017). It may be assumed that these subcortical abnormalities learning and emotional regulation, thereby contributing to the emo-
observed in preschool children decrease with age. However, a study on tional dysregulation and affective illness that affect many children with
the longitudinal neuroimaging of school-aged children and adolescents ADHD (Levy, 2004; Pliszka, 2000). The thalamus also receives projec-
with and without ADHD (Shaw et al., 2014) indicated that there was a tions from the cerebellum and connects reciprocally with the basal
similar trajectory between the striatal and pallidum volumes among ganglia, motor cortex, and premotor cortices to form a sensorimotor
individuals with and without ADHD. We also found an interaction be- network that drives the acquisition and implementation of learned
tween age and diagnosis in the bilateral pallidum. In the further ana- motor behaviors (Stepniewska et al., 2007). The delayed development
lysis, the pallidum in TD controls decreased with age, while that in of volumes in the thalamus could lead to gross and fine motor disorders,
ADHD individuals fit quadratic models and reached peak volume values as well as the slower processing speed recorded in children with ADHD
at ages 13.52 and 12.95 in left and right hemispheres, respectively. (Piek et al., 2007). However, we did not find the morphological dif-
Previous longitudinal studies (Raznahan et al., 2014; Wierenga et al., ferences in the thalamus in ADHD individuals compared with TD in-
2014), found a nonlinear developmental course of pallidum volume in dividuals, while we found that the thalamus fit a quadratic develop-
TD individuals, that reached peak volume values at 7.7–9.5 years of mental trajectory. Specifically, the thalamus volumes of both TD and
age. In our study, we observed only that the pallidum volume decreased ADHD individuals fit quadratic models with age, which reached peak
with age, which may have result from our sample age being from 7 to volumes at 11.33 and 9.91 in TD individuals but delays at 13.67 and
22 years of age, a range that was unable to capture the stage of volume 12.93 in ADHD individuals. We thought that the morphological ab-
rise and the peak volume value. In addition, due to the maturation normality of the thalamus in ADHD individuals that had been found
delay in ADHD, the trajectory fit an “inverted-U" developmental tra- previously may also result from the delay in maturation.
jectory. Overall, our results were consistent with the assumption that In addition, we found that the asymmetry of the thalamus was
subcortical basal ganglia changes are important factors in the patho- smaller in TD controls than in ADHD individuals, that is, TD controls
physiology of ADHD. had a stronger right lateralization, which was verified by the results of a
However, we found that the thalamus volume declined with age previous study (Wang et al., 2018). Furthermore, we found that the
only in TD youths, not in ADHD individuals. The thalamus projects to asymmetry of the thalamus decreased with age and we observed no
the frontal and parietal association cortices (Romanski et al., 1997), interaction between age and diagnosis. Previous studies suggest that
two thalamocortical pathways that support attentional processes. These attention directed to global aspects of visual processing is more right
circuits are important for stimulus seeking and distinguishing lateralized (Yamaguchi et al., 2000). ADHD and TD children are

Table 4
Fitting parameters for asymmetries of subcortical structures versus age in TD and ADHD individuals.
Parameter(p value)

Structure TD ADHD

Best fit model Intercept a

Age Age 2
Best fit model Intercepta Age Age2

Caudate None – – – None – – –

Putamen None – – – None – – –
Pallidum None – – – None – – –
Thalamus Linear −0.38(0.40) −0.09(0.03) – None – – –
Hippocampus None – – – Quadratic −10.74(4.83) 2.13(0.78) −0.08(0.03)
Amygdala None – – – None – – –

– = not applicable; n.s. = non-significant.

a
Intercept is the extrapolated value at age zero.

7
Y. Wang, et al. Progress in Neuropsychopharmacology & Biological Psychiatry 104 (2021) 110044

Fig. 3. The age trajectories of subcortical asymmetries, including the (A) thalamus, (B) hippocampus in TD (red color) and ADHD individuals (blue color). (For
interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

asymmetrical in space, but in the opposite direction. The lack of right 2004) may increase the risk of affective disorder in ADHD children and
lateralization in children with ADHD indicates that the right hemi- their family members (Peterson et al., 2001; Pliszka, 2000; Teicher
spheric attention network is damaged in individuals with ADHD. Pre- et al., 2002), even in family members without ADHD (Braaten et al.,
vious studies have investigated various aspects of brain structural 2003). We found a smaller volume of the amygdala in ADHD in-
asymmetry in ADHD (Hale et al., 2015; Shaw et al., 2009), which also dividuals than in TD controls, which was consistent with a portion of
reported increased asymmetry in the TD group. From the develop- previous works (Bonath et al., 2018; Frodl and Skokauskas, 2012;
mental trajectory in ADHD and TDC individuals, we found that the Hoogman et al., 2017; Sasayama et al., 2010) or a similar volume
maturation of the right thalamus delay is longer (left: 2.34; right: 3.01), (Amico et al., 2011, Perlov et al., 2008) in individuals with ADHD. The
whick may be lead to the right lateralization defect. amygdala fit a linearly increased model, and ADHD has a higher speed.
The hippocampus may play an auxiliary role in attention and cog- However, two studies have found that the amygdala also fit a curvi-
nition, and dysfunction in these is one of the characteristics of ADHD linear (quadratic or Poisson model) developmental model and reached
(Castellanos and Tannock, 2002; Pliszka, 2000; Sonuga-Barke, 2003). peaks at 19.7 and 22.6, respectively (Narvacan et al., 2017; Wierenga
We found no difference in the volume of the hippocampus between the et al., 2014). We hypothesized that this may result from the age range
ADHD and TDC groups. Previous studies have reported ambiguous re- of the study samples, which was less than 22 years and may not reach
sults for hippocampal volume changes in ADHD. Scholars have var- the peak value. Therefore, the amygdala in both ADHD and TD in-
iously reported increased volume (Plessen et al., 2006), decreased vo- dividuals fit an increased linear model, and due to the delay in ma-
lume (Bonath et al., 2018; Hoogman et al., 2017; Posner et al., 2014) or turation, the volumes in ADHD individuals have a steeper growth trend.
no difference (Amico et al., 2011; Perlov et al., 2008). With treatment Future studies could expand the age range to explore whether there was
and time, the changes from childhood to adulthood seem to gradually a peak volume in the amygdala at an older age.
decrease (Frodl and Skokauskas, 2012). In our study, the hippocampus
fit a quadratic model in ADHD individuals and TD controls, which
showed that hippocampal development in individuals with ADHD also 5. Limitations and future directions
lagged behind that of TDC individuals by several years. In addition, we
found that both the left and right hippocampus reached peak volumes There are some limitations in the present study that should be taken
at 16.75 and 15.53 years in TD individuals, respectively, but these into account. First, the data in the ADHD-200 sample database were
values were 19.17 and 16.90 years, respectively, in ADHD individuals. cross-sectional data for inferring developmental trajectories, therefore,
These results suggested that the development of hippocampal volume systematic cohort effects cannot be fully excluded. We have replicated
was asymmetric and lagged more in the left hemisphere, which was certain previously reported developmental trajectories in TD in-
confirmed in asymmetric development. We found that the development dividuals. Our results in ADHD individuals also need to be confirmed in
of hippocampus asymmetry also followed curvilinear trajectories, and future longitudinal studies. Second, previous studies have found
reached a peak at 13.59 in ADHD individuals but not in TD individuals, anomalous subcortical morphology between males and females.
which suggested that the dissymmetric development of the hippo- However, the percentage of females (75 females, 22.18%) in ADHD
campus may also result in ADHD. Previous studies have revealed that individuals was too low in our sample. Future studies could resolve the
the hippocampus was more asymmetric in ADHD individuals than in TD problem of sex percentage and further clarify the different develop-
individuals. In the present study, we found that the left hippocampus mental trajectories between ADHD and TD individuals in males and
was longer delayed than the right hippocampus, which suggested that females respectively. Third, because of the limited age range, we were
the damage to the left hippoccampus may be more serious. We found not able to define the age at which the adolescent phase of subcortical
that the peak of hippocampal asymmetry in maturation in ADHD in- volume growth levels off, transitioning into stable adult subcortical
dividuals was 13.59, which may be the low ebb of the development of dimensions. Perhaps the age of reaching this essentially static adult
verbal working memory. Although the results need to be verified by phase would also be later in the subjects with ADHD than in TD in-
future studies, they provide some meaningful information for the dividuals. In addition, we were not able to define the age at which the
treatment of ADHD children's vocabulary. amygdala reached peak volume. Finally, there were larger variances at
The pathophysiology of amygdala participation in ADHD (Levy, a later age point than early points because of fewer data points. Future
studies should investigate the developmental trajectories in lifespan

8
Y. Wang, et al. Progress in Neuropsychopharmacology & Biological Psychiatry 104 (2021) 110044

samples and confirm whether ADHD developmental delay will dis- References
appear at some point.
Alexander, G.E., DeLong, M.R., Strick, P.L., 1986. Parallel organization of functionally
segregated circuits linking basal ganglia and cortex. Annu. Rev. Neurosci. 9,
6. Conclusions
357–381.
Amico, F., Stauber, J., Koutsouleris, N., Frodl, T., 2011. Anterior cingulate cortex gray
In the present study, we found subcortical structures (caudate, pu- matter abnormalities in adults with attention deficit hyperactivity disorder: a voxel-
based morphometry study. Psychiatry Res. 191, 31–35.
tamen, pallidum, thalamus, hippocampus and amygdala) that followed
Asherson, P., Buitelaar, J., Faraone, S.V., Rohde, L.A., 2016. Adult attention-deficit hy-
curvilinear trajectories similar to those reported in previous studies. In peractivity disorder: key conceptual issues. Lancet Psychiatry 3, 568–578.
addition, we found that the volume of these subcortical structures in Barkley, R.A., 1997. Behavioral inhibition, sustained attention, and executive functions:
ADHD also delays the developmental trajectory, which suggested that constructing a unifying theory of ADHD. Psychol. Bull. 121, 65–94.
Bonath, B., Tegelbeckers, J., Wilke, M., Flechtner, H.H., Krauel, K., 2018. Regional Gray
ADHD may be characterized by a delay in subcortical maturation. The matter volume differences between adolescents with ADHD and typically developing
differences in subcortical volume in diagnosis and age between ADHD controls: further evidence for anterior cingulate involvement. J. Atten. Disord. 22,
and TD individuals may be the result of the age range of the selected 627–638.
Braaten B., E., Beiderman, J., Monuteaux C., M., Mick, E., Calhoun, E., Cattan, G., Faraone
sample. We conclude that the growth of the brain in individuals with V., S., 2003. Revisiting the association between attention-deficit hyperactivity dis-
ADHD is fundamentally different than in individuals with other neu- order and anxiety disorders: A familial risk analysis. Biol. Psychiatry 53 (1), 93–99.
rodevelopmental disorders; the patterns of growth differ greatly from Carrey, N., Bernier, D., Emms, M., Gunde, E., Sparkes, S., Macmaster, F.P., et al., 2012.
Smaller volumes of caudate nuclei in prepubertal children with ADHD: impact of age.
those in typically developing individuals. Instead, the present study J. Psychiatr. Res. 46, 1066–1072.
indicates that the maturation of brain architecture occurs in a highly Castellanos, F.X., Aoki, Y., 2016. Intrinsic functional connectivity in attention-deficit/
similar fashion between typically developing individuals and those with hyperactivity disorder: a science in development. Biol. Psychiatry Cogn. Neurosci.
Neuroimaging. 1, 253–261.
ADHD, with a significant delay in those with the disorder. This delay
Castellanos, F.X., Tannock, R., 2002. Neuroscience of attention-deficit/hyperactivity
also leads to a shift in which individuals with ADHD go through the disorder: the search for endophenotypes. Nat. Rev. Neurosci. 3, 617–628.
process of pruning the nerve connections that are part of the normal Castellanos, F.X., Lee, P.P., Sharp, W., Jeffries, N.O., Greenstein, D.K., Clasen, L.S., et al.,
2002. Developmental trajectories of brain volume abnormalities in children and
maturation process during adolescence. Further, we also found that the
adolescents with attention-deficit/hyperactivity disorder. JAMA. 288, 1740–1748.
asymmetric development of subcortical structures was abnormal in Child NSCO, 2008. ADHD Delays, Rather than Alters, Brain Development. Science Briefs.
ADHD individuals and resulted from the imbalance in the maturation National Scientific Council on the Developing Child.
delay of bilateral subcortical structures. The subcortical maturation Dickstein, S.G., Bannon, K., Xavier Castellanos, F., Milham, M.P., 2006. The neural cor-
relates of attention deficit hyperactivity disorder: an ALE meta-analysis. J. Child
delay may play an important role in the pathophysiology of ADHD. Our Psychol. Psychiatry 47, 1051–1062.
findings provide new potential targets to investigate in the pathophy- Eickhoff, S.B., Heim, S., Zilles, K., Amunts, K., 2009. A systems perspective on the ef-
siology of ADHD. fective connectivity of overt speech production. Philos. Trans. Series A Math. Phys.
Eng. Sci. 367, 2399–2421.
Ellison-Wright, I., Ellison-Wright, Z., Bullmore, E., 2008. Structural brain change in at-
Author contributions tention deficit hyperactivity disorder identified by meta-analysis. BMC Psychiatry.
8, 51.
El-Sayed, E., Larsson, J.O., Persson, H., Santosh, P., Rydelius, P.A., 2003. “Maturational
Yanpei Wang analyzed the data and wrote the draft of the paper. lag” hypothesis of attention deficit hyperactivity disorder: an update. Acta Paediatr.
Chenyi Zuo and Qinfang Xu amend and proofread the draft of the paper. 92, 776–784.
Qinfang Xu, Yuning Zhang, and Lei Hao participated in the discussion Fassbender, C., Schweitzer, J.B., 2006. Is there evidence for neural compensation in at-
tention deficit hyperactivity disorder? a review of the functional neuroimaging lit-
and offered some good ideas. All authors reviewed the manuscript. erature. Clin. Psychol. Rev. 26, 445–465.
Frodl, T., Skokauskas, N., 2012. Meta-analysis of structural MRI studies in children and
adults with attention deficit hyperactivity disorder indicates treatment effects. Acta
Ethical Statement
Psychiatr. Scand. 125, 114–126.
Garrett, A., Penniman, L., Epstein, J.N., Casey, B.J., Hinshaw, S.P., Glover, G., et al., 2008.
The present study was performed using publicly available data from Neuroanatomical abnormalities in adolescents with attention-deficit/hyperactivity
the ADHD-200 Consortium (http://fcon_1000.projects.nitrc.org/indi/ disorder. J. Am. Acad. Child Adolesc. Psychiatry 47, 1321–1328.
Grefkes, C., Eickhoff, S.B., Nowak, D.A., Dafotakis, M., Fink, G.R., 2008. Dynamic intra-
adhd200/). The ADHD-200 cohort contains both functional and ana- and interhemispheric interactions during unilateral and bilateral hand movements
tomical MRI data contributed by eight institutions. Each dataset was assessed with fMRI and DCM. Neuroimage. 41, 1382–1394.
approved by the research ethics review boards of each institution. Hale, T.S., Wiley, J.F., Smalley, S.L., Tung, K.L., Kaminsky, O., McGough, J.J., et al.,
2015. A parietal biomarker for ADHD liability: as predicted by the distributed effects
Signed informed consent was obtained from all participants or their perspective model of ADHD. Front. Psychiatry. 6, 63.
legal guardian before participation. Hart, H., Radua, J., Nakao, T., Mataix-Cols, D., Rubia, K., 2013. Meta-analysis of func-
tional magnetic resonance imaging studies of inhibition and attention in attention-
deficit/hyperactivity disorder: exploring task-specific, stimulant medication, and age
Declaration of Competing Interest effects. JAMA Psychiatry. 70, 185–198.
Hinshaw, S.P., Owens, E.B., Sami, N., Fargeon, S., 2006. Prospective follow-up of girls
with attention-deficit/hyperactivity disorder into adolescence: evidence for con-
The authors declare no competing financial or non-financial inter- tinuing cross-domain impairment. J. Consult. Clin. Psychol. 74, 489.
ests. Hobbs, M.J., Clarke, A.R., Barry, R.J., McCarthy, R., Selikowitz, M., 2007. EEG ab-
normalities in adolescent males with AD/HD. Clin. Neurophysiol. 118, 363–371.
Hoogman, M., Bralten, J., Hibar, D.P., Mennes, M., Zwiers, M.P., Schweren, L.S.J., et al.,
Acknowledgments 2017. Subcortical brain volume differences in participants with attention deficit
hyperactivity disorder in children and adults: a cross-sectional mega-analysis. Lancet
The authors acknowledge the contribution of ADHD-200 con- Psychiatry 4, 310–319.
Hsu, J.W., Lee, L.C., Chen, R.F., Yen, C.T., Chen, Y.S., Tsai, M.L., 2010. Striatal volume
sortium organizers for sharing the raw data. changes in a rat model of childhood attention-deficit/hyperactivity disorder.
This research was supported by Humanities and Social Science Fund Psychiatry Res. 179, 338–341.
of Ministry of Education of China (No. 18YJA90018), National Natural Kastner, S., O’Connor, D.H., Fukui, M.M., Fehd, H.M., Herwig, U., Pinsk, M.A., 2004.
Functional imaging of the human lateral geniculate nucleus and pulvinar. J.
Science Foundation of China (No. 31662083) and Natural Science Neurophysiol. 91, 438–448.
Foundation of Jiangsu Province (No. BK20191024). Levy, F., 2004. Synaptic gating and ADHD: a biological theory of comorbidity of ADHD
and anxiety. Neuropsychopharmacology. 29, 1589–1596.
Liddell, B.J., Brown, K.J., Kemp, A.H., Barton, M.J., Das, P., Peduto, A., et al., 2005. A
Appendix A. Supplementary data direct brainstem–amygdala–cortical ‘alarm’system for subliminal signals of fear.
Neuroimage. 24, 235–243.
Supplementary data to this article can be found online at https:// Mahone, E.M., Ranta, M.E., Crocetti, D., O’Brien, J., Kaufmann, W.E., Denckla, M.B.,
et al., 2011. Comprehensive examination of frontal regions in boys and girls with
doi.org/10.1016/j.pnpbp.2020.110044.

9
Y. Wang, et al. Progress in Neuropsychopharmacology & Biological Psychiatry 104 (2021) 110044

attention-deficit/hyperactivity disorder. J. Inter. Neuropsychol. Soc. 17, 1047–1057. Sasayama, D., Hayashida, A., Yamasue, H., Harada, Y., Kaneko, T., Kasai, K., et al., 2010.
Manjón, J.V., Coupe, P., 2016. volBrain: an online MRI brain volumetry system. Front. Neuroanatomical correlates of attention-deficit–hyperactivity disorder accounting for
Neuroinformatics. 10, 30. comorbid oppositional defiant disorder and conduct disorder. Psychiatry Clin.
Marcus, D.S., Harms, M.P., Snyder, A.Z., Jenkinson, M., Wilson, J.A., Glasser, M.F., et al., Neurosci. 64, 394–402.
2013. Human connectome project informatics: quality control, database services, and Shaw, P., Eckstrand, K., Sharp, W., Blumenthal, J., Lerch, J., Greenstein, D., et al., 2007.
data visualization. Neuroimage. 80, 202–219. Attention-deficit/hyperactivity disorder is characterized by a delay in cortical ma-
Marsh, A.A., Finger, E.C., Mitchell, D.G., Reid, M.E., Sims, C., Kosson, D.S., et al., 2008. turation. Proc. Natl. Acad. Sci. 104, 19649–19654.
Reduced amygdala response to fearful expressions in children and adolescents with Shaw, P., Lalonde, F., Lepage, C., Rabin, C., Eckstrand, K., Sharp, W., et al., 2009.
callous-unemotional traits and disruptive behavior disorders. Am. J. Psychiatry 165, Development of cortical asymmetry in typically developing children and its disrup-
712–720. tion in attention-deficit/hyperactivity disorder. Arch. Gen. Psychiatry 66, 888–896.
Mataro, M., Garcia-Sanchez, C., Junque, C., Estevez-Gonzalez, A., Pujol, J., 1997. Shaw, P., Malek, M., Watson, B., Sharp, W., Evans, A., Greenstein, D., 2012. Development
Magnetic resonance imaging measurement of the caudate nucleus in adolescents with of cortical surface area and gyrification in attention-deficit/hyperactivity disorder.
attention-deficit hyperactivity disorder and its relationship with neuropsychological Biol. Psychiatry 72, 191–197.
and behavioral measures. Arch. Neurol. 54, 963–968. Shaw, P., Malek, M., Watson, B., Greenstein, D., de Rossi, P., Sharp, W., 2013. Trajectories
Max, J.E., Fox, P.T., Lancaster, J.L., Kochunov, P., Mathews, K., Manes, F.F., et al., 2002. of cerebral cortical development in childhood and adolescence and adult attention-
Putamen lesions and the development of attention-deficit/hyperactivity symptoma- deficit/hyperactivity disorder. Biol. Psychiatry 74, 599–606.
tology. J. Am. Acad. Child Adolesc. Psychiatry 41, 563–571. Shaw, P., De Rossi, P., Watson, B., Wharton, A., Greenstein, D., Raznahan, A., et al., 2014.
McAlonan, K., Cavanaugh, J., Wurtz, R.H., 2008. Guarding the gateway to cortex with Mapping the development of the basal ganglia in children with attention-deficit/
attention in visual thalamus. Nature. 456, 391–394. hyperactivity disorder. J. Am. Acad. Child Adolesc. Psychiatry 53, 780–9 e11.
McFarland, N.R., Haber, S.N., 2002. Thalamic relay nuclei of the basal ganglia form both Soliva, J.C., Fauquet, J., Bielsa, A., Rovira, M., Carmona, S., Ramos-Quiroga, J.A., et al.,
reciprocal and nonreciprocal cortical connections, linking multiple frontal cortical 2010. Quantitative MR analysis of caudate abnormalities in pediatric ADHD: proposal
areas. J. Neurosci. 22, 8117–8132. for a diagnostic test. Psychiatry Res. 182, 238–243.
Mostofsky, S.H., Rimrodt, S.L., Schafer, J.G., Boyce, A., Goldberg, M.C., Pekar, J.J., et al., Sonuga-Barke, E.J., 2003. The dual pathway model of AD/HD: an elaboration of neuro-
2006. Atypical motor and sensory cortex activation in attention-deficit/hyperactivity developmental characteristics. Neurosci. Biobehav. Rev. 27, 593–604.
disorder: a functional magnetic resonance imaging study of simple sequential finger Sonuga-Barke, E.J., 2005. Causal models of attention-deficit/hyperactivity disorder: from
tapping. Biol. Psychiatry 59, 48–56. common simple deficits to multiple developmental pathways. Biol. Psychiatry 57,
Nakao, T., Radua, J., Rubia, K., Mataix-Cols, D., 2011. Gray matter volume abnormalities 1231–1238.
in ADHD: voxel-based meta-analysis exploring the effects of age and stimulant Sonuga-Barke, E.J., Halperin, J.M., 2010. Developmental phenotypes and causal path-
medication. Am. J. Psychiatry 168, 1154–1163. ways in attention deficit/hyperactivity disorder: potential targets for early inter-
Narvacan, K., Treit, S., Camicioli, R., Martin, W., Beaulieu, C., 2017. Evolution of deep vention? J. Child Psychol. Psychiatry 51, 368–389.
gray matter volume across the human lifespan. Hum. Brain Mapp. 38, 3771–3790. Stefanatos, G.A., Wasserstein, J., 2001. Attention deficit/hyperactivity disorder as a right
Perlov, E., Philipsen, A., Tebartz van Elst, L., Ebert, D., Henning, J., Maier, S., et al., 2008. hemisphere syndrome. Ann. N. Y. Acad. Sci. 931, 172–195.
Hippocampus and amygdala morphology in adults with attention-deficit hyper- Stepniewska, I., Preuss, T.M., Kaas, J.H., 2007. Thalamic connections of the dorsal and
activity disorder. J. Psychiatry Neurosci. 33, 509–515. ventral premotor areas in New World owl monkeys. Neuroscience. 147, 727–745.
Peterson, B.S., Pine, D.S., Cohen, P., Brook, J.S., 2001. Prospective, longitudinal study of Tamnes, C.K., Walhovd, K.B., Dale, A.M., Ostby, Y., Grydeland, H., Richardson, G., et al.,
tic, obsessive-compulsive, and attention-deficit/hyperactivity disorders in an epide- 2013. Brain development and aging: overlapping and unique patterns of change.
miological sample. J. Am. Acad. Child Adolesc. Psychiatry 40, 685–695. Neuroimage. 68, 63–74.
Piek, J.P., Dyck, M.J., Francis, M., Conwell, A., 2007. Working memory, processing speed, Teicher, M.H., Anderson, C.M., Polcari, A., Glod, C.A., Maas, L.C., Renshaw, P.F., 2000.
and set-shifting in children with developmental coordination disorder and attention- Functional deficits in basal ganglia of children with attention-deficit/hyperactivity
deficit–hyperactivity disorder. Dev. Med. Child Neurol. 49, 678–683. disorder shown with functional magnetic resonance imaging relaxometry. Nat. Med.
Pineda, D.A., Restrepo, M.A., Sarmiento, R.J., Gutierrez, J.E., Vargas, S.A., Quiroz, Y.T., 6, 470.
et al., 2002. Statistical analyses of structural magnetic resonance imaging of the head Teicher, M.H., Andersen, S.L., Polcari, A., Anderson, C.M., Navalta, C.P., 2002.
of the caudate nucleus in Colombian children with attention-deficit hyperactivity Developmental neurobiology of childhood stress and trauma. Psychiatr. Clin. N. Am.
disorder. J. Child Neurol. 17, 97–105. 25, 397–426.
Plessen, K.J., Bansal, R., Zhu, H., Whiteman, R., Amat, J., Quackenbush, G.A., et al., 2006. Toga, A.W., Thompson, P.M., 2003. Mapping brain asymmetry. Nat. Rev. Neurosci. 4,
Hippocampus and amygdala morphology in attention-deficit/hyperactivity disorder. 37–48.
Arch. Gen. Psychiatry 63, 795–807. Valera, E.M., Faraone, S.V., Murray, K.E., Seidman, L.J., 2007. Meta-analysis of structural
Pliszka, S.R., 2000. Patterns of psychiatric comorbidity with attention-deficit/hyper- imaging findings in attention-deficit/hyperactivity disorder. Biol. Psychiatry 61,
activity disorder. Child Adolesc. Psychiatr. Clin. N. Am. 9, 525–540 (vii). 1361–1369.
Pontius, A.A., 1973. Dysfunction patterns analogous to frontal lobe system and caudate Wagenmakers, E.J., Farrell, S., 2004. AIC model selection using Akaike weights. Psychon.
nucleus syndromes in some groups of minimal brain dysfunction. J. Am. Med. Wom. Bull. Rev. 11, 192–196.
Assoc. 26, 285–292. Wang, Y., Xu, Q., Li, S., Li, G., Zuo, C., Liao, S., et al., 2018. Gender differences in
Posner, J., Siciliano, F., Wang, Z., Liu, J., Sonuga-Barke, E., Greenhill, L., 2014. A mul- anomalous subcortical morphology for children with ADHD. Neurosci. Lett. 665,
timodal MRI study of the hippocampus in medication-naive children with ADHD: 176–181.
what connects ADHD and depression? Psychiatry Res. Neuroimaging 224, 112–118. Wang, Y., Tao, F., Zuo, C., Kanji, M., Hu, M., Wang, D., 2019. Disrupted resting frontal-
Pretus, C., Ramos-Quiroga, J.A., Richarte, V., Corrales, M., Picado, M., Carmona, S., et al., parietal attention network topology is associated with a clinical measure in children
2017. Time and psychostimulants: opposing long-term structural effects in the adult with attention-deficit/hyperactivity disorder. Front Psychiatry. 10, 300.
ADHD brain. A longitudinal MR study. Eur. Neuropsychopharmacol. 27, 1238–1247. Wang, Y., Zuo, C., Xu, Q., Liao, S., Kanji, M., Wang, D., 2020. Altered resting functional
Raznahan, A., Shaw, P.W., Lerch, J.P., Clasen, L.S., Greenstein, D., Berman, R., et al., network topology assessed using graph theory in youth with attention-deficit/hy-
2014. Longitudinal four-dimensional mapping of subcortical anatomy in human de- peractivity disorder. Prog. Neuro-Psychopharmacol. Biol. Psychiatry 98, 109796.
velopment. Proc. Natl. Acad. Sci. U. S. A. 111, 1592–1597. Wellington, T.M., Semrud-Clikeman, M., Gregory, A.L., Murphy, J.M., Lancaster, J.L.,
Romanski, L.M., Giguere, M., Bates, J.F., Goldman-Rakic, P.S., 1997. Topographic orga- 2006. Magnetic resonance imaging volumetric analysis of the putamen in children
nization of medial pulvinar connections with the prefrontal cortex in the rhesus with ADHD: combined type versus control. J. Atten. Disord. 10, 171–180.
monkey. J. Comp. Neurol. 379, 313–332. Wester, K., Irvine, D.R., Hugdahl, K., 2001. Auditory laterality and attentional deficits
Rosch, K.S., Crocetti, D., Hirabayashi, K., Denckla, M.B., Mostofsky, S.H., Mahone, E.M., after thalamic haemorrhage. J. Neurol. 248, 676–683.
2018. Reduced subcortical volumes among preschool-age girls and boys with ADHD. Wierenga, L., Langen, M., Ambrosino, S., van Dijk, S., Oranje, B., Durston, S., 2014.
Psychiatry Res. Neuroimaging 271, 67–74. Typical development of basal ganglia, hippocampus, amygdala and cerebellum from
Rowe, D.L., Robinson, P.A., Lazzaro, I.L., Powles, R.C., Gordon, E., Williams, L.M., 2005. age 7 to 24. Neuroimage. 96, 67–72.
Biophysical modeling of tonic cortical electrical activity in attention deficit hyper- Yamaguchi, S., Yamagata, S., Kobayashi, S., 2000. Cerebral asymmetry of the “top-down”
activity disorder. Inter. J. Neurosci. 115, 1273–1305. allocation of attention to global and local features. J. Neurosci. 20, RC72–RC.

10